DPC Holdings Limited (NYSE: DPC) — Third-Best Caster, Best-in-Class Multiple

An independent fundamental research note. Report date: 2026-07-25 · Price referenced: $49.91 (close 2026-07-24) · Market cap: ~$7.25B (145.2M shares per the prospectus; ~$7.45B if the over-allotment was exercised in full) · Net cash (pro-forma): ~$15M · EV: ~$7.2–7.3B Coverage status: Initiation of coverage. Stance (body): The analysis below carries no recommendation and no price target — it presents embedded expectations and scenario analysis only. The single exception is the opinion block immediately following.

⚡ Claude’s Take

This block is the author’s own subjective opinion and is offered as general information, not investment advice. It is the one place in this article where a position is taken; the detailed analysis that follows takes no position and carries no price target. Do your own work.

Verdict: AVOID here. A genuinely improved, structurally advantaged asset repriced in 21 trading days to roughly Howmet’s EV/EBITDA multiple on roughly half of Howmet’s margin — a ~5–6%-share #4 caster priced like a #1. Accumulation zone ~$27–34 (≈20–24x TTM EV/adj. EBITDA, ≈13–16x a credible FY2029 EBITDA), i.e. at or below the $33.00 IPO price and inside the $28–32 range these same underwriters marketed the deal at five weeks ago. Not a short — the demand backdrop is exceptional, the contracted revenue is real, and the largest shareholder was buying at $33.

The business is materially better than its GAAP accounts suggest, and the bear case should concede that first. Doncasters is one of only four globally scaled manufacturers of nickel- and cobalt-based aero and IGT castings; it is the only one that pairs casting with its own superalloy melt (three plants — Long Beach, Ross & Catherall in Sheffield, Bochum — 500+ specifications, sulphur below 0.5ppm); the barriers are real and slow (NADCAP/AS9100 over a year, OEM part-level qualification “often years,” 6–12 months to accredit an operator, long-lead furnaces); and roughly 40% of castings revenue is aftermarket on platforms that produce for ~30 years and spare for 20+. FY2025’s $(173)M net loss is largely the artefact of a 14% related-party PIK loan that no longer exists. The margin record is real: on the old Doncasters Group accounts, EBITDA margin went from ~7.6% in FY2018 to 16.5% in FY2025, and it came from gross profit — from the plants and from price — not from overhead leverage, since clean SG&A has been flat at ~11% of revenue throughout. Backlog moved $725M → $930M in one quarter. Four signed partnerships, including a 15-year agreement to enter aeroengine blades and vanes — Howmet’s most profitable franchise — should add “>$200M” of revenue with customers funding up to 80% of the capex. The UK defined-benefit pension, the classic 250-year-old-British-manufacturer landmine, has been entirely bought out. There is no factoring and no off-balance-sheet financing. None of that is the problem.

The problem is that the price pays for the destination and ignores the road. At $49.91 DPC is a ~$7.25B equity on ~48x trailing adjusted EBITDA, ~8.3x sales, 11.6x pro-forma book and 16x pro-forma tangible book — a higher EV/EBITDA multiple than Howmet, which earns 24.8% operating margins, ~21% ROIC and $1.4B of free cash flow, and which itself sits at roughly the 95th percentile of its own ten-year valuation history. DPC’s Engine Products segments earn 18.2% and 21.9% against Howmet Engine Products’ 33.3%. Post-tax ROIC is ~10–12% on adjusted EBIT — against a term loan that cost 10.8% and a lease discount rate of 13.0%; on GAAP EBIT FY2025 ROIC was negative. And the cash never arrives: operating cash flow was $(17)M in FY2024, $(7)M in Q1’26, and the +$42M in FY2025 came from stretching payables 20 days — strip the trade-payables move and it is ~$3M. On the $97M of already-committed capex and a working-capital build consistent with growth, pro-forma free cash flow is approximately zero. Pro-forma EPS is $0.37–0.49, i.e. 100–140x; the absolute floor multiple, torturing every assumption in management’s favour, is 66–75x. Then the governance: of ~$991M raised, $280M (28%) goes to nine people under a plan whose 13.5% compounding hurdle was deleted in March 2024 (“This led to a significant increase in the charge and liability”) and whose payout was then re-based to the $33 IPO price — CEO Quinn $85.8M, 66x his entire 2025 reported pay; another $155M repays a PIK loan to the shareholders who forgave 85% of it three months before pricing; and only ~$176M stays in the business. The company discloses three material weaknesses, including an admitted inability to verify that its IFRS-to-US-GAAP conversion adjustments “are complete and accurate,” with revenue recognised ex-works and no 404(b) attestation due for years.

The framing is a momentum-and-scarcity IPO trade at a late-stage capital-cycle top, not a quality compounder at a fair price. Only ~19–21% of the shares can trade; ~80% — over 3x the float — unlocks around 2026-12-21, held largely by distressed-credit funds (Searchlight, Mudrick, Corre, Bardin Hill, Hill City) whose cost basis is a defaulted 2020 loan. The +51% since pricing came with zero new company-reported information: no 10-Q, no 8-K, no call, no guidance. The stock has never traded within 30% of its IPO price; its all-time high of $53.50 is exactly Rothschild Redburn’s target, published on the first day the underwriters were permitted to publish. Meanwhile every Marathon warning light is on: gas-turbine prices up 195–300% in three years, 110 GW ordered against 60–70 GW of capacity, orders expected to peak in 2026, and Howmet ($470–500M capex), PCC, CPP, Chromalloy, the OEMs and the melters all adding capacity that lands 2028–2031 — including a customer-underwritten greenfield superalloy plant at DPC. The single most damning number in the file: while the OEMs’ selling prices roughly tripled, DPC’s adjusted EBITDA margin moved from 16.5% to 16.9%. That settles who holds the pricing power. Berkshire wrote down $9.8B on Precision Castparts in 2020 — “I was simply too optimistic about PCC’s normalized profit potential” — and GE wrote off $23B of Power goodwill in 2018 after the last gas-turbine boom. My base case, in which everything management promised happens on schedule, is ~$35–42. You are being asked to pay $49.91 for it.

Conviction: medium-high on the valuation, medium on the business. The single fact that would flip me bullish: two or three reported quarters showing Engine Products margins marching through the mid-20s with positive free cash flow after the MIP is paid, the capex step-up absorbed and the material weaknesses remediated — sustained margin convergence is the whole bull case and would make $50 look cheap in hindsight. The single fact that would flip me more bearish: the 2026-08-11 print revealing the Q1’26 margin was a mix or pass-through artefact, or evidence that customer capital contributions are netted against PP&E — which would make the 16.5% margin and the implied ROIC partly an accounting courtesy of the customers rather than an operating achievement. In fairness, two design details cut the other way and I weight them: the option ladder has escalating strikes ($33.00 / $36.30 / $39.93 / $43.92 / $48.31, average $40.29) requiring ~10%/yr compounding to pay — genuinely pro-shareholder — and J.F. Lehman put ~$143M of its own money in at $33.00 while selling nothing. Two dated events convert flow into information: 2026-08-11 (first reported quarter) and ~2026-12-21 (lock-up expiry). I would rather own this business at 20x than this multiple at any price. Tag: “Third-best caster, best-in-class multiple — and the customers own the capex.”

📈 Stock Price Action — IPO-to-Date Event Map

No five-year price history exists: DPC has been listed for 21 trading sessions (2026-06-25 to 2026-07-24). This block covers the entire trading record.

DPC priced its IPO at $33.00, above the $28.00–$32.00 range filed a week earlier, and upsized the deal ~19% to 27,858,585 shares (32,037,372 including the over-allotment option). The stock opened at $44.00 (+33.3% vs. the offer) and closed its first session at $46.88, +42.1%, on 18,650,400 shares — roughly two-thirds of the entire base deal traded on day one. It reached $53.50 intraday on 2026-06-30 (+62.1% vs. the offer), drifted to a lowest close of $45.06 on 2026-07-20, and finished 2026-07-24 at $49.91 — +51.2% versus the IPO price and 6.7% below its intraday high. There is no 52-week range and no meaningful moving-average, beta or drawdown history; realized volatility of daily closes is ~55.6% annualized on 20 observations. Critically, no company-reported financial information has been released since the prospectus — the first results as a public company are due 2026-08-11.

1. The deal was marketed at $28.00–$32.00 on 23.3M shares and printed at $33.00 on 27.9M — above range and upsized — then opened 33% higher still; the first-day tape is a fact about deal demand, not about the business. 2. The run to $53.50 came on no information: there was no filing, release or call between the 06-26 prospectus and 07-20. 3–4. The give-back and the flat, ultra-thin second week are consistent with ordinary post-IPO allocation turnover inside the period when dealers must still deliver a prospectus and no research is circulating. 5–6. The drift to $45.06 bottomed on the very day four syndicate banks published initiations and management set the results date — coverage arriving did not re-rate the stock, which argues the July decline was mechanical flow rather than a verdict on fundamentals. 7. The recovery to $49.91 tracked the aero-supplier complex almost tick-for-tick on GE Aerospace’s raised guidance, with DPC’s moves running roughly 2x the peer complex’s — the signature of sector beta transmitted through a ~21% float rather than anything company-specific.

1. Executive Summary

DPC Holdings Limited is the Jersey-incorporated, UK-headquartered holding company of the Doncasters Group — a ~248-year-old precision-engineering business that manufactures complex nickel- and cobalt-based investment castings and superalloys for aeroengines and industrial gas turbines (IGT), plus hot-side turbocharger wheels for internal-combustion and hybrid vehicles. It operates 14 principal plants across the UK, continental Europe, North America and Asia with 3,070 employees. It listed on the NYSE on 2026-06-25 at $33.00 a share, raising $919.3M of primary capital alongside a $68.76M insider placement and a $75M Qatar Investment Authority placement. Five weeks later it trades at $49.91, a market capitalisation of roughly $7.25B.

The business is real, structurally advantaged in one product niche, and materially subscale. DPC is one of four globally scaled players in a concentrated merchant market (HHI ~1,900–2,400) alongside Precision Castparts, Howmet and Consolidated Precision Products, and it holds roughly 5–8% of that market against #1 and #2’s combined 55–70%. Its distinctive asset is genuine upstream vertical integration — three superalloy melt shops supplying 100% of internal aero/IGT alloy demand, plus in-house ceramic cores at both large IGT plants. FY2025 revenue was $837M (IGT 42%, Aerospace 35%, Transportation 23%) with adjusted EBITDA of $138M (16.5%) and a GAAP net loss of $(173)M. Engine Products, the good half, earned $670M of revenue at a 20.3% margin; Turbo Wheels earned 6.5% on $185M.

The turnaround is real but smaller and more cyclical than presented. Management credits a post-2020 operational programme for revenue “more than doubling” from ~$365M in 2020 and margins rising from mid-single digits to 16.5%. Two qualifications matter. First, the 2020 base is close to a worst-possible starting point — the deepest aerospace downturn in history, the trough of the previous gas-turbine bust, and a group in a March-2020 debt-for-equity restructuring that wrote off ~£900M and transferred ownership from Dubai International Capital to its lenders. Second, as an emerging growth company DPC presents only two audited years; there is no disclosed revenue, margin or cash flow for 2021–2023 and no price/volume/mix/FX bridge anywhere, so the claim cannot be decomposed or verified. What is audited — FY2024 to FY2025 — shows genuine improvement: revenue +12.2%, gross margin 18.9% → 23.1%, and 33–46% segment drop-through driven by volume absorption and shortage-era repricing. Roughly $15–20M of the $91M revenue increase is currency translation, so organic growth was nearer 9.5–10%.

The demand backdrop is exceptional and externally verifiable — which is precisely the risk. GE Vernova is sold out through 2030 with a 116 GW backlog; Siemens Energy holds 60 GW firm plus 27 GW reserved; heavy-frame gas-turbine prices have risen 195–300% in three years on AI/data-centre electrification; CFM targets over 2,000 LEAP deliveries in 2026. DPC’s own backlog rose from $725M to $930M in a single quarter. But every Marathon capital-cycle top signal is present simultaneously: 110 GW of orders against 60–70 GW of capacity, orders expected to peak in 2026, and Howmet (capex guided $470–500M), PCC, CPP, Chromalloy, the OEMs themselves and the premium melters all adding capacity that becomes qualified capacity in 2028–2031 — the window in which DPC’s own contracted revenue arrives. The decisive evidence on who captures the scarcity rent is that the OEMs’ selling prices roughly tripled while DPC’s adjusted EBITDA margin moved 40 basis points.

Financial quality is the weakest part of the case. Post-tax return on invested capital on adjusted EBIT is ~10–12% (9.8% on gross invested capital), against a term loan priced at 10.8% and a lease discount rate of 13.0%; on GAAP EBIT, FY2025 ROIC was negative. Cash conversion is poor and deteriorating: DSO 58→68 days, DIO 88→110 days, cash conversion cycle 106→114 days, with receivables +31%, inventory +24% and payables +58% against revenue +12%. Operating cash flow was $(17)M in FY2024 and $(7)M in Q1’26; FY2025’s +$42M rests on a 20-day payables stretch and a $92M increase in an unpaid management-bonus accrual. On the $97M of already-committed capex and a working-capital build consistent with growth, pro-forma free cash flow is approximately zero. There is no factoring or off-balance-sheet financing — a genuine positive — and the UK defined-benefit pension has been fully bought out, leaving total pension liabilities of $26M.

Capital allocation and governance are poor. Of ~$991M of net proceeds, $330M repays a 10.8% term loan, $50M an ABL, $155M repays a shareholder PIK loan to the owners themselves (having forgiven 85% of it, ~$730M, effective 2026-03-19 and booked to additional paid-in capital), and $280M — 28% of the money raised — pays nine people under a management incentive plan whose 13.5% compounding hurdle was removed in March 2024, an amendment the filing says “led to a significant increase in the charge and liability.” Only ~$176M stays in the business. $121M of that compensation cost has never appeared in any income statement shown to IPO buyers. There is no returns-based metric in the go-forward compensation plan, the committee is chaired by a $23.9M plan beneficiary, and as an EGC there is no say-on-pay vote. Offsetting positives deserve weight: J.F. Lehman bought ~$143M of stock at $33.00, no shareholder sold a share in the offering, and the IPO option ladder carries escalating strikes averaging $40.29.

The valuation is the thesis. DPC trades at ~48x trailing adjusted EBITDA, ~8.3x sales, 16x pro-forma tangible book and roughly 100–140x pro-forma earnings — a higher EV/EBITDA multiple than Howmet, on half the margin, one-tenth the scale, from the #4 position, with negative GAAP earnings, no free cash flow, three disclosed material weaknesses in internal control, ex-works revenue recognition and 21 trading days of public history. Approximately 19–21% of the shares are freely tradeable; roughly 80%, held largely by distressed-credit funds, unlocks around 2026-12-21. To justify today’s enterprise value at the 20x multiple that mature aerospace suppliers command, DPC must earn ~$365M of EBITDA — 2.6x FY2025. The full contracted partnership programme adds ~$40M. A base case in which management delivers everything promised implies roughly $35–42 per share. This analysis takes no position; the scenarios in Section 10 are provided so a reader can judge the price rather than the company.

2. Business Overview

What DPC makes. DPC is a make-to-print manufacturer of safety-critical metal components that operate in the hottest, highest-pressure sections of jet engines and industrial gas turbines. The product set divides three ways.

Engine Products — Aerospace comprises structural castings for aeroengines: turbine centre frames, bearing housings, combustion diffusers, fin inducers, near-flow-path seals, blade outer seals, combustion seal segments, injector housings and nozzles. These are large, stationary, dimensionally exacting parts. DPC also supplies torque bars for landing-gear brakes and sells superalloy vacuum-induction-melted ingot to third parties. Content sits on CFM’s LEAP family (Boeing 737 and Airbus A320/A321), Pratt & Whitney’s GTF family, and GE Aerospace’s GEnx (Boeing 787), plus defence and space programmes.

Engine Products — IGT comprises turbine airfoils — the rotating blades and stationary vanes in the hot section — including large directionally solidified (DS) and single-crystal (SX) blades and vanes for heavy-frame turbines, an operation the company states is “limited to a small number of facilities globally,” plus heat tiles, seals and equiaxed structural castings. Content covers all categories of large and heavy-frame IGT platforms in production, notably Siemens Energy’s F-class and H-class machines.

Turbo Wheels manufactures hot-side turbocharger wheels for off-highway, commercial and passenger vehicles. Alloy for this segment is bought externally, principally from Chinese suppliers.

The manufacturing chain and where DPC is integrated. The core casting sequence runs wax assembly → ceramic shell → foundry (equiaxed, DS or SX pour) → post-cast processing (hot isostatic pressing, heat treatment, X-ray, non-destructive testing, dimensional inspection). DPC’s genuine integration is upstream: three superalloy melt shops — Long Beach (US), Ross & Catherall (Sheffield, UK) and Bochum (Germany) — supply 100% of internal aero/IGT alloy demand across more than 500 customised specifications with sulphur below 0.5 parts per million, and both large IGT plants run in-house ceramic-core units. Long Beach and Ross & Catherall also sell merchant alloy externally. Downstream integration is more qualified than the summary implies: the Business section says HIP, heat treatment, X-ray and NDT are performed in-house “at certain facilities,” while the risk factors state that “shell production, X-ray processing, hot isostatic pressing (HIP) and heat treatment … we outsource to third-parties,” with subcontractor concentration in HIP, X-ray, tooling and cores. A third-party X-ray processing incident disrupted DPC’s own output in 2025.

Segment economics. The reported 16.5% consolidated margin conceals a two-speed business:

Engine Products combined earned $670M of revenue at a 20.3% margin (FY2024: $576M at 16.3%), a 44.7% incremental margin. That 20.3%, not 16.5%, is the Howmet-comparable figure — and it remains roughly 13 points below Howmet Engine Products’ 33.3%. Note also that the unallocated corporate wedge is widening: $2M in Q1’25 to $5M in Q1’26, an annualised run rate moving from $10M toward ~$20M, a fresh ~120bp margin headwind running directly against the claim that margins will approach peers’. It is why 22.4% and 22.8% segment margins produced only 16.9% consolidated.

End markets. FY2025: Aerospace $291M (35%), IGT $351M (42%), Transportation $195M (23%). FY2024: Aerospace $267M (36%), IGT $280M (37%), Transportation $199M (27%). IGT delivered $71M of the $91M FY2025 revenue increase — 78% of the growth. This is a gas-turbine story marketed as an aerospace story, a distinction that matters because the two end markets have very different cyclical histories and because DPC’s aerospace exposure is essentially pure original-equipment.

Customers and concentration. Average customer tenure exceeds 20 years, and the named roster is the industry: GE Aerospace, Honeywell, Pratt & Whitney, Rolls-Royce, Safran, Ansaldo Energia, Doosan, GE Vernova and Siemens Energy. Concentration is high and rising: the top ten customers were 66% of FY2024 revenue, 68% of FY2025 and 70% of Q1’26; the top two were 34%, then 38% (Customer A 22%, Customer B 16%), then 43%. No single engine or turbine programme exceeds 7% of sales, which is a genuine mitigant — but nine programmes at one OEM is still one counterparty and one credit. Doncasters signed a contract worth over $1 billion with Siemens Energy in August 2023, which mechanically explains much of the IGT ramp and makes Siemens Energy the probable Customer A.

Contract structure. Approximately 70% of revenue sits under long-term agreements, typically five years or longer, with the remaining 30% on spot purchase orders. The LTAs were all renegotiated by the current management team since 2020 and now carry index-linked pass-through for metal, energy, labour, tariffs and general inflation — genuinely better protection than a typical Tier-2 supplier holds. But the LTAs are weaker than the headline reads: they “typically guarantee a minimum level of market share,” not volume (life-of-contract volume commitments exist in only three of the four strategic partnerships); “certain of our LTAs allow for termination by convenience by our customers”; certain LTAs “do not provide for minimum purchase commitments”; some are fixed-price; and three key contracts representing approximately 10% of revenue expire within one to two years. There is also an unreconciled contradiction in the filing: the 30% spot book allegedly earns “premium margins” while the LTA partnerships are simultaneously described as “margin accretive.”

Aftermarket mix — the largest quality gap versus Howmet. Roughly 60% of castings revenue is OEM and 40% aftermarket, but the aftermarket is almost entirely IGT, where airfoils are consumed and replaced across a ~20-year turbine life. Aerospace is essentially pure original equipment: DPC’s aero structural castings “typically last for the life of an engine.” Howmet’s spares mix went from 11% to 23% of revenue and is the principal driver of its margin expansion. DPC gets no meaningful aerospace aftermarket until 2029, when the 15-year blades-and-vanes partnership reaches run rate. Until then the aero business is a build-rate play, not an annuity.

Backlog. Firm purchase orders rose from $725M at 2025-12-31 to $930M at 2026-03-29, covering more than twelve months of aero and IGT casting production — an implied book-to-bill of roughly 2.1x on the Engine Products businesses and the single strongest hard datapoint in the filing. It is also almost certainly the partnership purchase orders landing in the same two concentrated customers.

Intellectual property — a structural ceiling. DPC manufactures to customer specification: “component designs provided by the customer. IP in the product design and tooling is owned by our customers.” Its only material patent estate is 34 Chinese patents covering turbocharger wheels. DPC holds no direct FAA production approvals of its own — it casts to OEM print under customer flow-down authority. The qualification barrier that protects the business is therefore OEM-granted, OEM-revocable, and the OEM captures the associated rent. This contrasts instructively with Chromalloy, which now holds its own FAA Parts Manufacturer Approval on CFM high-pressure-turbine blades.

Corporate history and structure. The listed entity was incorporated in Jersey as “Alloy Topco Limited” in November 2019 — the restructuring vehicle — and renamed DPC Holdings Limited in December 2025, converting to a Jersey public limited company immediately before the IPO. Dubai International Capital acquired Doncasters from Royal Bank of Scotland in 2006 for a reported ~£700M; creditors sought restructuring talks in 2019 over a $745M facility maturing April 2020; and in March 2020 ownership transferred to the first- and second-lien lenders in a scheme of arrangement that wrote off approximately £900M of a £1.22B funded-debt stack and installed a new chief executive and non-executive chairman. Doncasters had been shrinking before that: the Fasteners division was sold to Stanley Black & Decker for $426M (completed April 2018), two US Fabrications sites were disposed of in 2018–19, and Settas SA was sold to Precimetal in September 2019. The one acquisition since 2020 is Uni-Pol (2022), with no price, multiple or return disclosed — and Uni-Pol Mexico’s $9M of goodwill is now fully impaired, at the very site being converted into an aerospace hub. The disposal group held for sale is Ivostud, a German stud-welding unit inside Turbo Wheels with 169 employees, held for sale since 2024 and still unsold two years later on a letter of intent, written down $(9)M in FY2024 and written back +$5M in FY2025.

Verdict. A subscale, geographically fragmented, make-to-print precision-casting and captive-melt business with genuinely attractive content — heavy-frame IGT airfoils, aero structural castings on LEAP/GTF/GEnx, three superalloy plants — welded to a structurally inferior, ICE-exposed auto-parts business that consumes 24% of the assets and 28% of the workforce to produce 8% of segment profit. Customer concentration is high and rising, aerospace aftermarket exposure is effectively nil until 2029, and 70% of revenue sits under agreements that guarantee share rather than volume. The good half is genuinely good, and the reported 16.5% consolidated margin understates it (20.3%) while the segment margins overstate the enterprise. This is a real industrial business with a real niche — not the “scaled alternative to the two large industry participants” the prospectus describes.

3. Industry Dynamics

The value chain and where the profit sits. The chain runs from end customers (airlines and lessors; utilities, independent power producers and now hyperscalers) to platform OEMs (Boeing and Airbus; GE Aerospace, Safran/CFM, RTX-Pratt, Rolls-Royce and Honeywell in aero; GE Vernova, Siemens Energy, Mitsubishi Power, Ansaldo and Doosan in IGT), to Tier-1 casters (PCC, Howmet, CPP, DPC, Chromalloy, IHI, Aubert & Duval), to premium-melt superalloy suppliers (Carpenter’s SAO, ATI’s HPMC, Special Metals — owned by PCC — Aperam, Haynes, VDM, and DPC’s own three melt shops), and finally to raw nickel, cobalt, tantalum, tungsten, chromium, vanadium, ruthenium and hafnium plus revert scrap. DPC spans two adjacent nodes, melt and cast, which is genuinely unusual.

The economic profit does not sit evenly. Howmet’s Engine Products segment earned a 33.3% EBITDA margin in FY2025 and a record 36.6% in Q1’26 on $4.32B of revenue. DPC’s Engine Products earned 18.2% (North America) and 21.9% (Europe). Castings are an excellent business at scale, with aftermarket mix, and with yield discipline — industry analysis consistently identifies casting yield as the dominant profit lever, since a 70% versus 40% casting yield is all of the pricing and all of the profit. DPC captures roughly 55–60% of best-in-class unit economics, and its own prospectus concedes a “clear path … in-line with our best-in-class casting peer” — a 13-to-16-point aspiration against a rival five times its size that is simultaneously adding capacity.

The Precision Castparts precedent is the single most important analogue. Berkshire Hathaway paid ~$32.1B in equity (~$37B enterprise value) for PCC in 2016, explicitly for the installed-base economics of castings — and wrote down approximately $9.8B in Q2 2020, cutting roughly 10,000 jobs (−30%) from the forging, casting and fastener business off a 33,417 base. Buffett’s explanation was precise: “I was simply too optimistic about PCC’s normalized profit potential.” The moat never broke. What broke was normalised-earnings optimism at a build-rate peak, compounded by fixed-cost absorption in a downturn. All three failure modes apply to DPC with more force: a thinner margin cushion, less diversification, and a far higher entry multiple. It is worth noting that PCC has since recovered — 2024 revenue $10.4B, pre-tax earnings ~$1.9B, its first year above the 2019 peak — so the “weakened competitor” premise underpinning DPC’s share-gain story is now stale.

Commercial aerospace: strong and accelerating, mid-cycle on original equipment. The 737 MAX has moved from 38 per month to an FAA-cleared 47 per month (May 2026), targeting 53 by end-2026 with ~70 under study; the A320 family is moving from ~65 toward 75 per month by mid-2027. CFM delivered more than 1,000 LEAP engines in the first half of 2026 (Q1 +63% to 520) and is tracking above 2,000 for the year, targeting 2,500 by 2028. Airbus and Boeing each hold seven-to-eight-year backlogs, more than 15,000 aircraft are on order, and the installed fleet is ~15 years old, a thirty-year high. The Pratt & Whitney GTF powder-metal recall is, counterintuitively, a net positive for a hot-section caster — roughly 3,000 engines to inspect, 600–700 removals and 835 aircraft grounded at end-October 2025 (~38% of A320neos) suppresses new-build demand but massively lifts high-margin spare-part consumption.

Against that backdrop, one number is uncomfortable: DPC’s Aerospace revenue grew only +9.0% in FY2025 ($267M → $291M) while LEAP output ramped hard and Howmet guided commercial aero up 15% for FY2026. DPC attributes the shortfall to a temporary production shutdown at Groton to install new equipment. Q1’26 then jumped +43% ($65M → $93M), but off a weak comparator ($65M against a $73M FY2025 quarterly average). The share-gain claim in aerospace currently rests on a single quarter.

IGT: the growth vector and the danger. This is where the story is. GE Vernova’s gas-turbine backlog reached 116 GW in Q2 2026 (from 80 GW at end-2025), targeting 125 GW by year-end, sold out through 2030 with more than half of 2031 already contracted; output moves from a 20 GW run rate in Q3 2026 to 24 GW in 2028. Siemens Energy holds 60 GW firm plus 27 GW reserved and is booked out to FY2028 with four-year customer waits. Mitsubishi’s backlog is ~$35.6B. Prices have risen 195% on Wood Mackenzie’s measure (toward ~$600/kW by end-2027, roughly 3x 2019), with GE Vernova’s first-half 2026 orders priced 10–20 points higher per kilowatt than Q4 2025; press reports put the three-year increase at 300%. DPC’s IGT revenue grew +25.4% in FY2025 and +29% in Q1’26, and roughly 40% of castings revenue is aftermarket weighted to IGT — a genuine annuity on a 20-year airfoil replacement cycle. Two caveats on the disclosure: the prospectus’s two load-bearing IGT market claims (a ~9% CAGR to 2035 and a “>170GW” combined OEM backlog) are unattributed, and the latter is already stale.

Transportation: already declining. $199M in FY2024 to $195M in FY2025, −2.0%, and flat at $50M in Q1’26, at a 6.5% margin falling to 3.7%. Published turbocharger forecasts of 7–9% CAGRs are contradicted by DPC’s own numbers, which are the better evidence. A 3–5% annual structural decline costs little EBITDA but removes the low-capex cash currently helping fund the aero/IGT programme and caps consolidated growth by roughly 0.7–1.2 points a year.

Market size and DPC’s actual share. Building bottom-up from disclosed and estimated competitor revenue — Howmet Engine Products castings ~$2.4–2.8B, PCC Investment Cast ~$2.2–2.7B, CPP ~$0.4–0.6B, Chromalloy new-make ~$0.3–0.5B, DPC $642M, others ~$0.7–1.2B — the merchant nickel/cobalt aero-and-IGT castings market is roughly $7–9B, and adding premium nickel/cobalt melt at $3–4.5B gives a served market of about $10–13B. That puts DPC at 5–6% share, with #1 and #2 holding 55–70% between them. The estimate carries ±30% uncertainty and excludes OEM in-house casting (GE, Rolls-Royce Rotherham, Siemens Energy, Mitsubishi Takasago), which is both unquantifiable and itself a competitive risk. Third-party market-research TAMs were discarded as internally contradictory — one widely-cited figure puts “superalloy investment castings” at $1.04B, smaller than Howmet’s Engine Products segment alone.

The capital cycle — the central industry judgement. Applying Marathon’s supply-side lens, every precondition for a capital-cycle top is present in IGT castings simultaneously: record demand (110 GW ordered against 60–70 GW of capacity), record prices (+195% to +300% in three years), record backlogs, publicly announced multi-year sold-out positions, universal announced capacity expansion, and — the classic tell — new equity capital entering the supply side.

The supply response, quantified: Howmet capex has gone $199M (2021) → $453M (FY2025), guided to ~$470–500M for FY2026 with ~70% into Engine Products, and management states demand-and-capital negotiations are finalised with six of seven IGT customers. PCC is expanding Ravenswood, West Virginia explicitly to enable “capacity-based pricing,” and on 2026-03-01 completed its first acquisition since the Berkshire takeover — the UK’s Morvern Group (wax patterns, ceramic core assemblies and tooling; ~£22M revenue), based in Derby and Worcester, which is to say in DPC’s own back yard. CPP (Warburg Pincus and Berkshire Partners) is building a new facility in Euclid, Ohio. Chromalloy has entered the highest-value aftermarket airfoil pool with its own FAA PMA and a US airline launch customer. The OEMs are adding directly: GE Vernova from 55 to 70–80 heavy-duty units a year, Mitsubishi doubling by FY2030, Siemens Energy adding 61,000 square feet of blade-and-vane capacity. Premium melt is adding too: Carpenter’s ~$400M Athens VIM project commissions in FY2028 with qualifications running to ~2030, and ATI’s nickel remelt comes online in Q4 2026 with VIM melt in 2027.

Three observations follow, and they are the sharpest findings in this article.

First, DPC’s fourth partnership underwrites a customer-funded greenfield superalloy plant. The bull case for premium melt has rested on the claim that no new greenfield premium-melt capacity has been built in the West in decades and that all additions are brownfield and slow to qualify. DPC’s own filing describes three agreements with an aerospace OEM “underwriting greenfield superalloy expansion,” beginning production in 2029. Whether or not DPC’s specific project is large, its existence is evidence that the supply-side discipline that thesis relies on is breaking — and that OEM cheques are what is breaking it.

Second, customer-funded capacity is how a capital cycle turns, not how a moat is built. DPC presents “up to 80% of customer contribution to capital investment” as proof of entrenchment. Read from the supply side it is the opposite. The OEMs face a shortage they cannot fix quickly through two incumbent suppliers, so they are subsidising the expansion of a third and fourth source. DPC’s own prospectus says the quiet part out loud: the partnerships provide “additional capacity and competition within these supply-constrained markets.” The purpose of a supply-side response is to lower the price of supply. DPC will itself be an incumbent facing that competition by 2029. The risk transfer is structural: the OEM funded up to 80% of the asset, holds volume commitments as options (with convenience-termination rights in certain LTAs), and can re-source; DPC staffs the plant, powers it, depreciates it and absorbs the fixed cost when the shortage clears.

Third — and this is the empirical test of the first two — the OEMs’ selling prices roughly tripled while DPC’s adjusted EBITDA margin moved from 16.5% to 16.9%. Forty basis points. If DPC held real pricing power in a market where its customers’ output prices trebled, it would appear here. It does not.

Where regulation genuinely distorts the cycle — the honest bull case. Qualification slows the supply response materially and is the reason this cycle can run longer than a normal industrial one. NADCAP/AS9100 accreditation plus baseline casting capability takes over a year and substantial capital; OEM part-level process approval “often tak[es] years” and is specified down to individual process parameters; critical quality roles require 6–12 months of operator training; and new melt furnaces qualify grade-by-grade over years. Net: capacity announced in 2026 does not become sellable qualified capacity until roughly 2028–2031. That lag is the moat’s real content, and it is why returns are unlikely to break in 2026 or 2027. Two mechanisms defeat it over time: the additions are being pre-qualified now at customer expense, compressing the usual friction; and the demand peak arrives first, with Wood Mackenzie expecting gas-turbine orders to peak in 2026 and GE Vernova’s own contracted 2030–31 position meaning incremental order flow must decelerate even if deliveries do not.

Competitive intensity. Four globally scaled casters, HHI roughly 1,900–2,400 — concentrated but not a duopoly, and concentration is falling at the margin because the OEMs are deliberately qualifying third and fourth sources. Buyer power is extreme and worsening (top two at 43% of Q1’26 revenue, termination-for-convenience clauses, OEM-owned design and tooling, explicit dual-source/re-source/in-source risk, and no DPC-held FAA production approval). Supplier power is also worsening: cobalt entered 2026 at ~$56,414/tonne, the highest since July 2022, after the DRC capped hydroxide exports at ~96,000 tonnes for 2026; Indonesia, 60.2% of global nickel output, cut its 2026 mining quota to 260–270Mt from 379Mt; and Turbo Wheels’ automotive-grade superalloys are “primarily sourced from suppliers in China.” Index-linked pass-through is genuinely better-than-typical protection, but it is a lagged timing hedge (the filing concedes “a delay between an increase in our costs and our ability to increase the prices of our products”), it covers only ~70% of revenue, and it mechanically dilutes percentage margin when input prices rise — the same surcharge distortion documented in the Carpenter work. Cobalt at a four-year high is a live FY2026 margin-optics risk.

Substitution, assessed seriously. Ceramic matrix composites are the real long-run threat. GE’s GE9X, entering service in 2026, carries more than 100 individual CMC parts across five hot-section components: combustor inner and outer liners, high-pressure-turbine Stage 1 shrouds and nozzles, and HPT Stage 2 nozzles. HPT shroud manufacturing has reached Manufacturing Readiness Level 10 — full-rate production. GE has demonstrated the first rotating CMC component and invested over $105M incrementally in CMC manufacturing. Every one of those parts is precisely the static hot-section nickel content DPC casts. The 424B4 contains zero mentions of “composite” or “CMC” — a conspicuous disclosure gap. Timing honesty matters here: this is a 2030s–2040s threat to next-generation platforms, not to LEAP/GTF/GEnx or F- and H-class content, which will be produced and spared for decades. It does not impair the base case; it caps the terminal value, and it is the reason a business like this should not carry a secular-growth multiple. Additive manufacturing is not yet a substitute for DS/SX airfoils — the literature is consistent that AM creep properties “are not yet good enough to replace single-crystal directionally solidified investment castings,” and Howmet’s own management has explicitly rejected additive for single-crystal rotating parts. AM’s real roles today are repair, complex internal cooling features and printed moulds. DPC’s prospectus mentions 3D printing once, generically.

Regulation, tariffs and energy. Tariffs are currently a non-event to mild positive: UK civil aircraft parts and jet engines are zero-rated under the WTO Civil Aircraft Agreement per the US–UK framework, and civil aircraft and parts are excepted from the Section 232 steel, aluminium and copper tariffs for the UK, EU, Japan and Korea. The 2026-07-09 Section 232 proclamation on aircraft, engines and parts imposes no new tariffs but mandates negotiations within 180 days — roughly January 2027, a live binary. Turbo wheels are not covered by the civil-aircraft carve-out and face the automobile-parts regime. Energy is a structural handicap that is currently non-binding: UK industrial electricity runs ~28–32p/kWh, with supported industrial prices around £86/MWh against ~£69 in France and ~£60 in Germany — roughly 40% above France and Germany and about 4x US levels — and DPC’s highest-margin segment sits in the worst geography (Chard, Deritend and Ross & Catherall in the UK, Bochum in Germany). Yet Engine Products — Europe out-earns North America, so mix and pass-through more than offset it today. The tail risk is an energy shock, which is exactly what a lagged pass-through does not cover; the India plant lost two weeks in March 2026 to LPG unavailability tied to the Middle East conflict, taking five weeks to return to full rate. Add CSDDD and CSRD, conflict-minerals rules (cobalt/DRC, tantalum and tungsten), CERCLA and the UK Contaminated Land Regime on legacy foundry sites, and ITAR/EAR/OFAC controls on defence and space content.

Cyclicality and the crucial precedent. DPC’s FY2020 trough was $365M. On the old Doncasters Group accounts, FY2018 revenue was ~£612.7M with ~£46.6M of EBITDA (7.6%), so FY2025’s $837M is only modestly above FY2018 in absolute terms and the true peak-to-trough experience was roughly 55%, not the gentler picture the prospectus conveys. That comparison must be caveated — the historical perimeter included since-divested units and FY2018/19 on the current perimeter is not disclosed, which is the single most important missing number in the file — but the margin improvement from 7.6% to 16.5% does look real.

The gas-turbine bust of 2015–2020 is the analogy to draw explicitly. In 2017 Siemens estimated large gas-turbine demand falling to ~110 units a year against ~400 units a year of world capacity — roughly 4x overcapacity. Siemens Power & Gas orders fell 41%, revenue 11% and profit 23% “due to overcapacities,” with 6,900 job cuts and European plant closures. GE wrote off $23B of Power goodwill in 2018 — more than it had paid for Alstom’s power assets three years earlier — and cut its dividend to a penny; John Flannery’s post-mortem was that “we were too optimistic for too long about the market.” Today the same three OEMs are sold out, prices have tripled, and OEMs, casters, private-equity owners and now the public market are all adding capacity into an order book expected to peak this year. What partially rescues the present cycle: the driver is AI and data-centre electrification with better-capitalised counterparties; the ~40% aftermarket mix on a 20-year-consumption installed base is a genuine annuity that did not protect the OEMs but does protect a spares caster; and qualification lags delay the overbuild to 2028–2031. What rhymes: in both cycles the marginal unit was priced by scarcity, and scarcity pricing is not a durable valuation input. Note finally that aero and IGT are less diversifying than they look — they share superalloy supply, DS/SX capacity, foundry labour and, in 2026, a common capex-cycle driver.

Verdict — two distinct answers. Is this a structurally good industry? Qualified yes at the top; only marginally so for DPC’s position within it. The structural case is real: qualification incumbency on platforms that produce for 30 years and spare for 20+, genuine concentration, multi-year entry lags, a ~40% annuity aftermarket, and melt-plus-cast integration that is hard to replicate. Against it: extreme and rising buyer power, worsening cobalt and nickel supplier concentration, a demonstrated ~55% peak-to-trough, a European energy handicap, and CMC already in full-rate production on DPC’s exact static hot-section content. The decisive evidence on where DPC sits within the good industry is its own P&L — 16.5% consolidated and ~20% in Engine Products against Howmet Engine Products’ 33.3% — and the cleanest falsification of the moat claim is Turbo Wheels, where DPC calls itself “the market leader and one of the only worldwide suppliers” and earns 6.5%.

Is it a good industry at this point in the cycle? No — this is a late-cycle entry point, and the evidence is unusually unambiguous. Capacity contracted in 2026 arrives 2028–2031 into an order book expected to peak in 2026. That is precisely the configuration that produced Siemens’ 4x overcapacity, GE’s $23B write-down and Berkshire’s $9.8B PCC write-down the last two times around. The genuine mitigants — the aftermarket annuity, index pass-throughs, contracted volumes in three of four partnerships, a backlog up 28% in a quarter, and the 2028–2031 qualification lag — mean the cycle should not break in 2026 or 2027. They do not change the direction of travel.

4. Competitive Position

Naming the moat in Greenwald’s taxonomy. The framework recognises three genuine advantages — supply/cost, demand/customer captivity, and economies of scale combined with captivity — and DPC’s position against each is instructive.

Supply or cost advantage: absent. DPC earns a 16.5% consolidated EBITDA margin against Howmet’s ~31%, and 12.7% adjusted operating margin against Howmet’s 24.8% — literally half. Its best segment, Engine Products — Europe at 21.9%, sits roughly 1,150 basis points below Howmet Engine Products. Fourteen plants averaging $60M of revenue each is the definition of subscale. A company with a cost advantage does not earn half its rival’s margin.

Customer captivity: present, but not proprietary. The lock-in is real — part-level qualification, customer-owned tooling, 6–12 months to accredit an operator, NADCAP/AS9100 taking over a year, OEM approval taking years — but it attaches to whoever is qualified, and PCC, Howmet and CPP hold it identically on far more parts. It describes the industry, not DPC’s position within it. Worse, because DPC holds no FAA production approval of its own and the customer owns the design and the tooling, the captivity asset sits on the customer’s side of the table.

Economies of scale plus captivity: DPC is on the wrong side. This is decisive, because it is the one form of advantage that actually governs this industry. DPC is roughly 8% of PCC’s revenue and 10% of Howmet’s. Greenwald’s prescription for the subscale player is to dominate a defensible niche. DPC is instead doing the opposite: head-to-head expansion into Howmet’s core franchise, with customer-funded DS/SX aero blades at Oxford.

Where a real niche moat probably does exist. Large heavy-frame DS/SX IGT airfoils. The installed manufacturing base for these is genuinely scarce — DPC states the ability to make them is “limited to a small number of facilities globally” — and it is where the money is: Engine Products — Europe at 21.9%, up 540 basis points, on 46% drop-through, underwritten by a billion-dollar Siemens Energy contract. If there is a moat in this business, it lives here, and it is narrow.

The market-share-stability test cannot be run — and that is itself the finding. Greenwald’s cleanest empirical test of a moat is share stability. DPC never quantifies its market share, its addressable market, or its competitive position anywhere in 832,000 characters of prospectus. Contrast Howmet, which discloses more than 50% global gas-turbine-blade share, or ATI, which discloses sole-source positions on five of seven advanced nickel grades. The only explicit leadership claim DPC makes anywhere is for the 6.5%-margin turbocharger-wheel business. And DPC’s central thesis is that it is taking share — “winning majority roles for parts and volume from competitors.” Under Greenwald, share instability is evidence against barriers. DPC cannot simultaneously claim high barriers to entry and material share gains without conceding that the barriers are lower than advertised, or more precisely that its customers have chosen to lower them.

Why it is gaining share — three causes, none of which is a DPC advantage. First, PCC withdrew capacity: Berkshire’s $9.8B write-down and roughly 10,000 job cuts (−30% of the forging, casting and fastener workforce) in the first half of 2020, with revenue only regaining its 2019 peak in 2024 — while DPC deliberately retained labour through the trough. Second, the OEMs are deliberately dual-sourcing, in DPC’s own words to create “additional capacity and competition,” because “customers are actively diversifying their supply chains.” Third, there is a shortage: castings are the acknowledged binding constraint across a large OEM backlog with three-to-eight-year lead times. Scarcity rent is not a moat, and two of these three causes are reversible.

The decisive internal experiment: Turbo Wheels. DPC claims to be “the market leader and one of the only worldwide suppliers of hot-side turbocharger wheels” — and earns 6.5%, falling to 3.7% in Q1’26, on declining revenue. Same company, same vacuum casting technology, same alloy expertise, same “operational toolbox,” same claimed global leadership — roughly fifteen points less margin than Engine Products — Europe. The differences are that this end market is not in shortage and the buyers are BorgWarner, Cummins and Garrett rather than Siemens Energy. If the advantage lived in DPC’s casting capability, it would show up here. It does not. Margin in this business is set by end-market scarcity and customer power, not by DPC’s manufacturing skill. This is the most powerful single piece of evidence in the entire competitive analysis, and it is internal to the company.

Vertical integration is a supply-security asset, not a cost advantage — and the peer set proves it. The claim is that owning three melt shops “allows us to capture additional profit in the casting value chain.” Three facts undercut it. PCC is more integrated than DPC — it bought Special Metals, the leading producer of nickel-based superalloys, in 2005–06, and TIMET for $2.9B in 2012 — and PCC produced the $9.8B write-down. Howmet is not integrated into nickel melt at all: its own 10-K states it “requires nickel, titanium, aluminum, cobalt, and superalloy materials,” purchased on the open market — and Howmet earns the industry’s best margins and ~21% ROIC. And the merchant melters out-earn DPC’s castings: Carpenter’s SAO segment hit a record 35.6% ex-surcharge margin with 17.2% ROIC, and ATI’s HPMC earns 23.6%. Meanwhile DPC never discloses superalloy revenue, margin or transfer price, so the value-capture claim is unverifiable. Integration is genuinely valuable as insurance in a melt-constrained market — the external superalloy market is capacity-short and DPC is immune to it — but the evidence does not support it as a source of superior returns.

Customer-funded capacity: bond or lien? This is the analytical fork on which the moat verdict turns. The facts: “up to 80% of customer contribution to capital investment”; a 15-year aerospace blades-and-vanes agreement, 7-year and 9-year IGT extensions with guaranteed volumes, and three agreements underwriting greenfield superalloy; “>$200M” of incremental revenue with full-year effect in 2029; terms “in line with our typical LTAs” plus capacity reservation payments; and capacity “allocated to that customer only,” reusable only “subject to the specific terms and conditions of the agreements and approval and tooling requirements.”

The bond reading is that customer capital in your plant is a fifteen-year commitment no rival can match, pre-qualifies capacity at someone else’s expense, and signals that the OEM regards DPC as strategically necessary. That is real and should be credited.

The lien reading is stronger as disclosed, for eight reasons. Funded capacity is reserved to a single customer, so DPC forfeits the option to sell it into the 30% spot book it says earns “premium margins” — customer funding converts optional high-margin capacity into contracted capacity. Reuse requires the funder’s consent. Pricing is fixed-with-index, which is cost-plus: a customer that paid 80% of the plant will not also pay a return on it. No pricing mechanism, return, asset title or accounting treatment is disclosed anywhere. Concentration rises to 43%. Williamson’s hold-up logic applies — the holder of a single-customer-specific asset is the weaker party at renewal. The framework rests on LTAs carrying convenience-termination rights. And DPC itself says the purpose is to give OEMs “competition.” Moreover, DPC’s uniqueness claim is false as stated: it asserts it is “the only competitor … able to provide OEMs with dedicated capacity,” yet ATI discloses capex that is “partly customer-funded.” Customer-funded capacity is an industry-wide shortage response, not a DPC innovation.

The ROIC test: pass, with a heavy qualification. Computing invested capital bottom-up gives $641M on a net basis and $810M on the harsher gross basis (gross PP&E, gross intangibles, goodwill, net working capital and right-of-use assets). Post-tax at the 25% UK statutory rate: on adjusted EBIT of $106M, FY2025 ROIC was 12.4% on net invested capital and 9.8% on gross. On GAAP EBIT it was negative. Levelizing the management incentive plan as the recurring cost it economically represents takes it to 6.6% and 5.2%. The relevant hurdle is not abstract: DPC’s own term loan cost 10.8% and its lease discount rate is 13.0%. So the honest statement is that DPC’s returns on capital are somewhere between 5% and 12% depending on how you treat management’s pay, and even the most generous defensible figure sits at or barely above its own cost of debt. The ladder against peers is clear: Howmet ~21% > Carpenter 17.2% > DPC ~10–12% > ATI 13.7% on comparable adjusted bases — and DPC’s figure is measured at a cyclical peak, on a 2020 acquisition-accounting asset basis, before the coming stock-compensation charge. The trend is unambiguously improving on every definition, and the Q1’26 annualised figures are the first that plausibly clear a hurdle rate; but one annualised quarter is not a track record.

The cleanest moat-width discriminator in the whole file: Howmet stayed profitable through COVID. DPC was in a distressed debt-for-equity restructuring with mid-single-digit EBITDA margins. A moat is what protects you in the bad year.

Turbo Wheels is a value-destroying capital sink, not the “key cash generator” it is described as. Segment assets are $218M — 24.4% of non-corporate assets — consuming 28.2% of the workforce and 21.6% of revenue to produce 8.1% of segment EBITDA. Strip $6M of depreciation from $12M of EBITDA and it earns roughly $6M of EBIT on $218M of assets: a 2.8% pre-tax return on assets, while carrying $28M of goodwill. At an auto-supplier 6x it is worth perhaps $70–75M against $218M of carried assets — real impairment exposure. Divesting it is the economically correct action; Ivostud, held for sale for two years and still unsold, is the warning about execution.

How much of the turnaround is skill? The disclosure does not permit a clean decomposition, and that is the finding: two audited years, no price/volume/mix/FX bridge anywhere, and the words “on-time delivery,” “scrap,” “yield” and “utilization” appearing repeatedly without a single quantification in any period. What can be said about the $365M → $837M move: the base was collapsed and distressed (2020 was ~9.5 months of COVID plus a scheme of arrangement); there is an unquantified inorganic contribution (Uni-Pol, 2022); the IGT/data-centre boom and one $1B contract explain the largest single increment; the post-COVID aero recovery contributed less than it should have (+9%, held back by Groton); price is real, management-attributable and shortage-enabled (all key LTAs repriced since 2020); FX is an unquantified tailwind with no constant-currency disclosure despite 47.5% of revenue billed in Europe and the UK; and there is at least $6M of allocation noise, since Turbo Wheels’ 140bp “improvement” came from a corporate recharge credit falling from $8M to $2M, explicitly “a partial credit of certain costs recharged in 2024.” Against that, the genuine execution is not trivial: $170M of capex, a doubled Groton shell line, a new Ross & Catherall VIM furnace, the Deritend consolidation, an RTX Platinum Award, labour retained while PCC cut 30%, and four customer-funded partnerships no peer publicly matches at this scale.

The honest split: the FY2024→FY2025 350bp margin gain is mostly operating leverage on IGT volume plus shortage-era repricing, with modest real productivity, an unquantified FX tailwind and allocation noise. The five-year doubling is mostly cycle and recovery off a distressed base. That is good opportunistic management of a cyclical asset, not evidence of structural advantage — and the filing lacks the data to prove otherwise. One relevant credibility note cuts in management’s favour: CEO Michael Quinn was a PCC Group Vice President from 2005 to 2014, and COO Jason Mays spent 1990–2006 at Howmet. DPC’s operating leadership was hired out of the two companies it is now competing with. That is the most credible mechanism behind the turnaround — and it is key-person risk rather than an institutional moat.

Verdict: a narrow, partly borrowed competitive advantage — real in one product niche, absent at the enterprise level. In Greenwald’s taxonomy DPC has no supply/cost advantage, no proprietary customer captivity, and is on the wrong side of economies of scale plus captivity. What it does have is a genuine niche position in large heavy-frame DS/SX IGT airfoils where the manufacturing base is scarce and the margins show it; a very large slice of cyclical scarcity rent, which is not a moat; and a position as the third source that four OEM customers have decided they want to exist and are funding — an advantage granted by customers, exercisable at the customers’ price, and revocable by them. The moat-to-financial-outcome test is met only conditionally: the margin that would deteriorate without it is Engine Products — Europe’s. And the internal natural experiment of Turbo Wheels is strong evidence that in this business margin is set by end-market scarcity and customer power rather than by DPC’s manufacturing capability. This is a good, hard business in an excellent cycle from a weak relative position, with a credible but unbuilt 2029 plan to improve that position using its customers’ capital on terms it has not disclosed. Not a durable advantage; more than no advantage. Weak-to-narrow.

5. Growth History and Forward Opportunities

The historical record, and its limits. DPC discloses two audited years. FY2024 revenue was $746M; FY2025 was $837M, +12.2%. Q1’26 was $237M against $188M, +26.1%. Beyond that, the record is a single unaudited assertion — “approximately $365 million revenue in 2020” — with nothing at all for 2021, 2022 or 2023. Emerging-growth-company relief permits the two-year presentation, and DPC has taken it. The consequence is that the headline growth claim cannot be trend-checked or decomposed by an outside investor, and there is no price/volume/mix/FX bridge anywhere in the document.

Growth by end market:

Three things stand out. First, this is an IGT growth story: gas turbines produced 78% of the FY2025 increase, and the segment carrying that exposure (Engine Products — Europe) is also the highest-margin one. Second, aerospace underperformed its own end market — +9.0% in a year when LEAP output ramped hard and Howmet guided commercial aero +15%. Management attributes this to a temporary production shutdown at Groton to install new capacity, which is credible, but the Q1’26 +43% rebound comes off a weak $65M comparator against a $73M FY2025 quarterly average, so the aerospace share-gain claim currently rests on one quarter. Third, Transportation is already in decline, and DPC’s own numbers are better evidence than the published 7–9% turbocharger CAGRs.

Organic versus acquired versus currency. There has been one acquisition since 2020 — Uni-Pol in 2022 — with no disclosed price, multiple or contribution, and Uni-Pol Mexico’s $9M of goodwill is now fully impaired. Currency is the larger unquantified factor: DPC reports in US dollars against a sterling and euro cost base, with 47.5% of revenue billed in Europe and the UK and roughly 5% appreciation in both currencies against the dollar across 2025. On a reasonable estimate, $15–20M — some 17–22% — of the $91M revenue increase was translation, putting organic growth nearer 9.5–10% than 12.2%. Because costs translate up alongside revenue the effect on margin percentage is roughly neutral, so this does not explain the margin gain; but it does deflate the growth headline, and no constant-currency disclosure is provided anywhere.

The margin trajectory is not the clean upward march the summary implies. Quarterly adjusted EBITDA margin through FY2025 ran 15.4%, 18.4%, 16.3%, 15.9%, then 16.9% in Q1’26. Q2 2025 was the peak. Gross margin did improve genuinely and progressively — 18.9% (FY24) to 23.1% (FY25) to 24.1% (Q1’26) — and that is the right place for it to come from, since clean SG&A excluding the incentive plan and IPO costs has been essentially flat at 10.9%, 11.1%, 11.2% and 10.1% of revenue. There is no overhead-leverage story here; all of the improvement is in the plants and in price.

And the operating leverage is already decelerating. The FY2024→FY2025 incremental gross margin was a striking 57% and the incremental adjusted EBITDA margin 45%. In Q1’26 against Q1’25 those figures were 31% and 22% — roughly half. Segment drop-through fell from 46% to 33% in Europe and 39% to 33% in North America. This is exactly what one expects as easy fixed-cost absorption is consumed and new capacity brings its own depreciation and overhead. A 30–35% forward drop-through is the defensible assumption, which on 10% revenue growth implies roughly +100–130 basis points of annual adjusted EBITDA margin. That is real and valuable — and it points toward the mid-teens by 2028–29, not toward Howmet’s mid-20s.

Forward opportunities — the four strategic customer partnerships. These are the substance of the growth case and they are genuine contracted business, not pipeline:

1. A 15-year agreement with a major aerospace OEM to supply aeroengine blades and vanes, including customer investment to significantly increase capacity at two aerospace plants, applying DPC’s DS and single-crystal casting technology to both OEM and aftermarket applications.
2. A 7-year extension with a major IGT OEM for blades and vanes, with expanded scope, increased duration, significant capacity investment and guaranteed volumes.
3. A 9-year extension with a major IGT OEM for guaranteed volumes of directionally solidified turbine airfoils, with investment to increase IGT capacity.
4. Three agreements with an existing major aerospace OEM increasing casting and superalloy volumes and margins “while underwriting greenfield superalloy expansion.”

Together these are expected to deliver “more than $200 million” of incremental annual revenue at full run rate, with “up to 80% of customer contribution to capital investment.” Timing is the critical and under-appreciated detail: the programme “is expected to reach full run rate in 2028 with the full year effect recognized in 2029,” and the greenfield superalloy facility “is expected to begin production in 2029.”

Item 1 is the most strategically significant fact in the filing. DPC’s aerospace business today is structural castings — stationary parts that last the life of the engine and generate almost no aftermarket. Entering aero blades and vanes puts DPC into the consumable, high-margin, aftermarket-rich product category that is the engine of Howmet’s economics, whose spares mix went from 11% to 23% of revenue. If it works, it addresses DPC’s single largest quality gap. It is also a direct assault on Howmet’s core franchise, funded by a customer that wants a second source — which tells you something about the terms on which it will be priced.

Sizing the opportunity honestly. More than $200M of revenue at a 20% incremental margin — generous for volume-guarantee contracts negotiated by a counterparty funding 80% of the plant — is roughly $40M of incremental EBITDA on a FY2025 base of $138M. Layer on organic growth from the narrowbody ramp, the IGT installed-base aftermarket and continued repricing, and a credible FY2029 picture is roughly $1.3B of revenue at 19–20% margins, or about $254M of EBITDA. That is genuine, substantial growth — an 84% increase in EBITDA over four years — and it is the reason this is not a short.

The three constraints on that path. First, timing collides with the capital cycle: the revenue arrives in 2028–29, precisely the window in which the industry’s aggregate qualified capacity additions land, and one year after gas-turbine orders are expected to peak. Second, the capex is in front of the revenue and it is DPC’s own commitment: committed-but-unspent capital expenditure went $11M → $53M → $97M in fifteen months against guidance of ~$58M over the next twelve, taking capex from 3.7% of revenue toward 7–12%; and not one dollar of the promised customer funding is quantified anywhere in the financial statements. Third, concentration deepens: the partnerships are with the same handful of customers whose share of revenue has gone from 34% to 43% in two years. The growth model does not diversify the business; it concentrates it.

Verdict: high-quality growth in composition, lower-quality in origin, and back-end loaded. The composition is good — contracted, multi-year, aftermarket-extending, in genuinely capacity-short end markets, with volume guarantees in three of four agreements. The origin is weaker: the FY2025 growth was 78% gas turbines in a market whose prices have tripled, roughly a fifth of it was currency, aerospace lagged its own end market, and the five-year “doubling” is largely recovery from a distressed 2020 base that cannot be verified. And the incremental economics are decelerating in real time, from 45% drop-through to 22%. The forward opportunity is real, it is contracted, and it does not arrive until 2028–29 — which is both the reason to be interested in the business and the reason not to pay 48x trailing EBITDA for it today.

6. Financial Quality

The reported statements.

Start with the line the prospectus does not print. DPC’s income statement runs straight from gross profit to interest expense; there is no operating-income line anywhere in the filing. Compute it and the result is uncomfortable: GAAP operating income went backwards, from +$31M (4.2% margin) in FY2024 to $(5)M (−0.6%) in FY2025. The $88M increase in SG&A — $58M of incremental management-incentive charge plus $18M of IPO costs — more than consumed the $52M gross-profit gain. An equity story sold on “margin progression” reported a GAAP operating loss in its headline year, and chose a presentation format in which that fact does not appear. Q1’26 recovered to +$12M (5.1%).

Adjusted EBITDA and what it excludes. The company’s $138M FY2025 figure adds back interest, tax and D&A ($221M — definitional and legitimate), $18M of IPO costs (legitimate), $9M of debt-modification and disposal-group items (legitimate and symmetric), a $16M FX gain (legitimate, and conservative since it removes a gain, including a one-off non-cash gain on restructuring the intra-group financing) — and three items that do not belong. Site-closure and refinancing costs are recurring for a serial restructurer. Claims, settlements and litigation are an ordinary cost of running a foundry. And $87M of long-term management incentive plan expense is added back to a margin measure while $280M of it is being paid in cash.

The right way to handle the incentive plan is not to ignore it. It was implemented in March 2020 and paid in June 2026 — 6.29 years — so $280M / 6.29 = $44.5M a year of genuine incentive cost. On that levelized basis, FY2025 adjusted EBITDA is $88.5M, a 10.6% margin, not 16.5%; FY2024 is $52.5M (7.0%); and Q1’26 annualised is $115.5M (12.2%). On any honest reading DPC is a ~10–12% EBITDA-margin business today — which is the difference between “approaching Howmet” and “a subscale caster.” (For valuation purposes the $280M is a one-time claim already settled out of IPO proceeds and therefore already reflected in the post-IPO capital structure used in Section 10; it is not double-counted there. But it should not be excluded from any judgement about the historical earning power of the business, and the going-forward replacement is $30–40M a year of stock compensation.)

Returns on capital — the framework’s central test, and DPC fails it or barely passes. Invested capital computed bottom-up is $641M on a net basis and $810M on the harsher gross basis (gross PP&E, gross intangibles, goodwill, net working capital, right-of-use assets). At the 25% UK statutory rate:

The relevant hurdle is not theoretical. DPC’s own term loan carried a 10.8% effective rate and its lease discount rate is 13.0%. So post-tax returns on capital are somewhere between 5% and 12% depending on how management’s pay is treated, and even the most generous defensible figure sits at or barely above its own cost of debt. Against peers on comparable adjusted bases: Howmet ~21% > Carpenter 17.2% > ATI 13.7% > DPC ~10–12% — and DPC’s figure is struck at a cyclical peak, on a 2020 acquisition-accounting asset basis, before the coming stock compensation. The trend is unambiguously improving on every definition, and the Q1’26 annualised figures (14.6% net / 11.6% gross) are the first that plausibly clear a hurdle rate. One annualised quarter is not a track record. Return on equity is meaningless pre-IPO — shareholders’ equity was $(790)M, $(964)M and $(237)M at the three balance-sheet dates — and pro-forma post-IPO equity of $625M implies ROE of roughly 9–12%.

Cash conversion is poor and deteriorating — and FY2025’s positive figure was bought.

Receivables grew 31.1%, inventory 24.0% and payables 58.2% in FY2025 — against revenue up 12.2% and cost of sales up 6.4%. Every working-capital line grew at a multiple of the underlying flow. Of the +$70M “payables, accrued expenses and other liabilities” inflow that produced the positive FY2025 operating cash flow, trade payables alone contributed +$39M — a twenty-day stretch — and “accruals and deferred income” went from $35M to $75M, including accrued-but-unpaid IPO costs (only $5M of offering expenses had actually been paid by 2026-03-29). Strip the trade-payables stretch alone and FY2025 operating cash flow falls from $42M to roughly $3M. The management-incentive accrual grew $92M in the same year and was also unpaid. Q1’26 operating cash flow was $(7)M, which the company attributes to a $43M inventory build on rising input metal costs — meaning operating cash flow was negative in the very quarter whose numbers were used to price the IPO.

Conversion ratios tell the story plainly: operating cash flow as a percentage of adjusted EBITDA was 30% in FY2025, (18)% in FY2024 and (18)% in Q1’26. Free cash flow was $(52)M, +$11M and $(17)M — cumulatively negative $58M across three periods. There is no evidence in the disclosed record that this business converts EBITDA into cash.

Two genuine positives on the working-capital question deserve to be stated. There is no factoring, no securitisation and no supply-chain finance — explicit searches return zero hits and the company states it engages in no off-balance-sheet financing and holds no variable-interest entities. The DSO deterioration is real and unmasked, which is more than many industrials can say. But the Wells Fargo ABL (£90M / ~$121M, secured on receivables and inventory, SONIA/SOFR/EURIBOR + 3.0%, maturing July 2027) functions as a daily receivables sweep — $1,006M drawn against $1,019M repaid in FY2025 through a $121M facility — and its drawn balance was $40M at 2024-12-31, $1M at 2025-12-31, and $19M again eleven weeks later at 2026-03-29. Paying it down to $1M at the FY2025 year-end flattered the $524M adjusted net debt figure and the 3.8x leverage headline that appears in the prospectus summary.

The FY2025 tax credit is entirely non-cash and roughly two-thirds one-off. On a $(204)M pre-tax loss the tax line was a $31M benefit, comprising a $9M current-tax expense and a $(40)M deferred benefit. The specific driver is disclosed: “$23.8 million deferred tax asset has been recognized in the period in relation to historic losses in the UK due to the implementation of restructuring,” which the MD&A describes as “revisions to valuation allowances of $21 million.” The valuation allowance churned $28M of additions against $34M of releases for a net $(6)M rate-reconciliation line — considerable estimation latitude in the year that set the offering price. Of the $20M year-over-year improvement in reported net loss, the tax line contributed $39M of swing while GAAP operating income fell $36M. Cash taxes paid were $6M. Going forward, the book effective rate should trend toward the UK’s 25% statutory rate while the cash rate stays near 10–15% for years, shielded by $219M of UK non-trading losses, $105M of German losses and $184M of combined interest carryforwards, none expiring.

The balance sheet is now genuinely clean — but not because the business earned it. Total debt fell from $1,434M at 2025-12-31 to $712M at 2026-03-29 after the PIK forgiveness, and pro-forma post-IPO borrowings are $195M against $210M of cash — approximately $15M of net cash, with interest expense collapsing from $222M to $5–9M net. That is a real and substantial improvement worth roughly $39M a year of cash interest to new shareholders. The mechanism, however, was an $774M related-party debt forgiveness credited directly to additional paid-in capital plus $535M of debt repaid from public money. Shareholders’ equity went from $(964)M to +$625M without the business earning a dollar of it. Note also that the company’s headline “3.8x” leverage excluded the PIK loan entirely; including it, net debt to adjusted EBITDA at 2025-12-31 was 10.2x.

One specific structural item that appears unaddressed. The Term Loan matures April 2030 “provided that the maturity date under the Shareholder PIK Loan … is extended to July 23, 2030, otherwise the maturity date of the Term Loan will be December 5, 2027.” DPC is repaying the PIK loan from IPO proceeds, not extending it. Whether that satisfies, breaches or moots the condition is not discussed anywhere in the filing. It is a small residual balance and DPC is net cash, so this is not a solvency question — but a potential springing maturity inside eighteen months on the remaining term debt is the kind of item that should not require a reader to find it.

Pensions — the expected landmine is not there, and this is the cleanest thing on the balance sheet. For a ~248-year-old UK manufacturer one expects a legacy defined-benefit scheme with a deficit and a trustee holding leverage over corporate events. Instead: “the company’s legacy defined benefit pension plans, comprising the Triplex and Doncasters schemes, have been entirely de-risked and fully bought out by insurance companies, with all associated liabilities transferred. As a result, the Group has no remaining defined benefit obligation.” There is no UK scheme, no deficit, no trustee, and therefore no UK Pensions Regulator leverage over the IPO or over future capital allocation. What remains is trivial: US funded status $(7)M, unfunded US and German obligations $(15)M, and a German early-retirement plan at $(2)M — $26M in total, 0.36% of market capitalisation, with employer contributions of $1M a year.

The asset base — I tested the harvesting hypothesis and it is not supported. Gross PP&E is $337M against $221M net, so the base is only 34.4% depreciated — mid-life, not worn out — with depreciation of $23M implying a ~14.7-year average life consistent with stated policy. Critically, capex/depreciation was 1.35x in FY2025 and 1.64x in FY2024: DPC has been reinvesting above depreciation, so the reported EBITDA is not being harvested out of a starved plant. That should be said plainly rather than assumed away. Two qualifications remain. Cumulative capex of “more than $170 million since 2020” is roughly $28M a year against ~$22–23M of annual depreciation, so six years of “strategic expansion” produced only $5–6M a year of net asset growth — the plant has been maintained, not expanded; the expansion begins now. And gross PP&E to revenue is only 0.40x, roughly half Howmet’s, which either reflects genuinely high asset productivity or means the carrying basis — set by acquisition accounting around the 2020 restructuring — materially understates replacement cost, in which case true maintenance capex is above book depreciation. The $97M of committed capex, the greenfield plant, the admission that they “replaced aged machinery with state of the art equipment,” and the Groton installation that took production offline in 2025 all point toward the second reading.

Customer-funded capex is quantified nowhere. This is the single largest analytical gap in the file. The prospectus promises “up to 80% of customer contribution to capital investment” across four partnerships, yet there is no accounting policy for customer contributions to property, plant and equipment, no deferred-income or contract-liability balance attributable to them (deferred revenue is $14M and contract liabilities are stated not to be material), no reduction-of-PP&E disclosure, and no line in either investing or operating cash flow. Not one dollar is quantified. The consequence matters directly: if contributions are recorded as deferred revenue they become a multi-year revenue and margin drag as they amortise; if they are netted against PP&E they permanently depress the depreciation charge and inflate reported ROIC on an understated asset base, which would make part of the “path to peer margins” an accounting artefact rather than an operating achievement.

Other earnings-quality flags. Revenue is recognised largely ex-works: “control of the goods transfers to the customer when the products are made available for collection at the Group’s premises” — goods need never leave the plant to become revenue. The scrap-return contra-revenue estimate roughly doubled, with the refund liability going $8M → $17M and the matching right-of-return asset $6M → $13M. Intangible amortisation is scheduled at “$8 million in fiscal year 2027, $8 million in 2028, and $23 million in fiscal years 2029, 2030 and 2031” — a near-tripling on a $96M net balance with no explanation and no FY2026 figure given. And no stock-compensation expense, grant-date fair value or valuation assumption is disclosed anywhere for the 7.5M options, despite roughly $98M of estimated grant-date fair value implying $30–40M a year of coming charges that no adjusted measure will show.

Internal control. DPC and KPMG together identified three material weaknesses as of 2025-12-31: no formal control framework including general IT controls; “lack of resource and expertise necessary to apply U.S. GAAP conversions from IFRS, including whether the list of adjustments are complete and accurate”; and insufficient segregation of duties over the review and approval of journal entries across components. Remediation explicitly names “revenue recognition at our manufacturing sites.” No Section 404 evaluation has been performed and, as an emerging growth company, no auditor attestation is required for up to two more years. There is no going-concern issue — the statements are prepared on a going-concern basis with no substantial-doubt language, and that should be said clearly. The auditor is KPMG LLP, Birmingham, serving since 2021, with a dual-dated report (2026-04-14, except Note 21 at 2026-06-15). Taken together with ex-works recognition, a doubled return estimate, $62M of valuation-allowance churn and a Level-3 valuation on the incentive plan, every reported figure in this article should be treated as provisional.

Verdict: mixed-to-poor, with a genuinely improving operating core buried under a badly-disclosed capital structure and incentive arrangement. The operating improvement is real and should not be dismissed: gross margin 18.9% → 23.1% → 24.1%, 33–46% segment drop-through from volume absorption and genuine pricing in a capacity-short market, flat clean overhead, UK pension risk permanently extinguished, no factoring or off-balance-sheet financing, and a post-IPO balance sheet that is genuinely net cash. But GAAP operating income went backwards; operating cash flow was negative in two of three periods and the one positive year was bought with a payables stretch; three-period cumulative free cash flow is negative $58M; post-tax ROIC of 5–10% on gross invested capital sits below any plausible cost of capital; and the honest EBITDA margin is ~10.6%, not 16.5%. Do economics improve with scale? Yes — but less than reported, from a lower base, and already decelerating from 45% drop-through to 22%. The honest destination is mid-teens margins by 2028–29, not Howmet’s mid-20s.

7. Capital Allocation

Sources and uses — the crux of the section.

$435M — 43.9% of net proceeds — went to insiders and existing owners. Only $176M stayed in the business, 37% less than the management bonus. FY2025 capex was $31M and committed capex is $97M, so the retained amount is genuinely useful; but the growth capacity that the equity story rests on is funded by customers, not by this raise. The honest characterisation is that public shareholders funded a management payout and a balance-sheet repair.

The management incentive plan. A cash plan established as part of the March 2020 restructuring, with payouts geared to repayments of the Shareholder PIK Loan above defined thresholds — so the act of listing is the trigger. Nine participants share $280M, an average of $31M each. The disclosed splits: CEO Michael Quinn $85.8M, CFO David Egan $38.1M, COO Jason Mays $25.4M, and — notably — two non-executive directors, Dirkson Charles $29.9M and Nicholas Sanders $23.9M. Those five take $203.1M, or 72.5%; four unnamed participants share the remaining ~$76.9M.

Three features make this worse than its size alone.

The hurdle was lowered twice, both times in management’s favour. The thresholds originally “increased by 13.5% on a quarterly basis,” and the note states plainly: “The rules of the plan were updated in March 2024 such that the 13.5% compounding was removed. This led to a significant increase in the charge and liability.” The plan was then amended again at the IPO to fix payouts to the $33.00 offering price, with approximately 1,250,000 immediately-vested “MIP Recognition Grant” options — no performance condition — awarded as consideration for agreeing to the amendment, worth ~$21M of intrinsic value at today’s price. Three amendments, each favouring the participants.

The CEO sat on the committee that set his own award. The filing discloses that “the members of our then Nominating and Remuneration Committee consisted of Mr. Charles, Mr. Sanders and Mr. Quinn. Mr. Quinn was and is our CEO” — and all three were plan beneficiaries. The go-forward compensation committee is chaired by Nicholas Sanders, a $23.9M beneficiary.

A large part of the cost never reaches an income statement. The liability was $146M at 2025-12-31 and $159M at 2026-03-29 (discounted at 25%); the payment is $280M. $121M — 43% of the largest cash cost in the company’s history — goes straight to accumulated deficit without passing through any P&L shown to IPO buyers. And the $87M that did pass through was added back to adjusted EBITDA. For scale: $280M is 28.3% of net proceeds, 2.03x FY2025 adjusted EBITDA, 33.5% of FY2025 revenue, and $1.93 per share. Meanwhile the FY2025 Summary Compensation Table shows CEO total pay of $1,294,155 with zero in the non-equity-incentive column — in the year of the $87M charge, three months before an $85.8M payday. As an emerging growth company DPC holds no say-on-pay vote.

Equity incentives — where the analysis cuts the other way, and it matters. The 13,812,500-share reserve is 9.52% of shares outstanding, including 7,499,995 options granted at the IPO. The initial premise that these were all struck at $33.00 is wrong, and the correction is materially favourable to shareholders: the 6,249,995 IPO Grants come in five equal tranches with escalating strike prices of $33.00 / $36.30 / $39.93 / $43.92 / $48.31, cliff-vesting in years one through five — an average strike of $40.29 requiring roughly 10% annual compounding before the top tranches pay anything. Intrinsic value at $49.91 is about $81M, not the ~$127M a flat-$33 assumption implies. This is genuinely pro-shareholder design and the one unambiguously good element of the compensation architecture. Allocations: Quinn 1,674,144, Egan 1,004,485, Mays 803,589, and 104,146 to each of six directors; C. Alexander Harman, the J.F. Lehman nominee, received nothing — appropriate for a sponsor designee. The offsetting red flag: the 1.25M-a-year grant cap disappears entirely once the stock trades above $66 (twice the IPO price) after June 2027.

Incentive alignment is otherwise absent. The go-forward annual bonus is 100% discretionary. There is no ROIC metric, no EBITDA target, no revenue target, no relative-TSR measure and no performance share units anywhere in the disclosed plan. For a business whose expansion is funded by customers, whose returns on capital sit at roughly the cost of its own debt, and whose value depends on whether new capacity earns an adequate return, the absence of any returns-based measure is a material governance failure. Alignment has been created as cash out, not equity in: pre-IPO, insiders owned just 0.44% of the company, four of nine directors owned zero shares, and the CEO owned 27,729 shares.

Ownership — a 2020 lender consortium heading for the exit.

This is a who’s-who of distressed-debt and special-situations credit funds — precisely what a leveraged loan converting to equity produces. Their cost basis is a defaulted 2020 claim, not $33.00, and their exit is the public market. Combined 5%-holder ownership is roughly 53.6% against a free float of about 19–21%.

A crucial disclosure point on the “existing shareholders’ contribution.” The prospectus dilution table credits existing shareholders with total consideration of $774M at an average $6.85 per share, inviting the reader to conclude they invested $774M of cash. They did not. Additional paid-in capital was $0 at 2025-12-31; the Q1’26 equity statement shows a single line, “Capital contribution 774.” That $774M is the non-cash credit from forgiving their own PIK loan in March 2026 (the arithmetic reconciles: $878M grown at 14% for 2.7 months ≈ $911M × 85% = $774M, with the residual 15% being the $137M carrying value). The 2020 lenders contributed approximately zero cash. The accounting treatment — crediting APIC rather than recognising a P&L gain — is correct under ASC 470-50-40-2 for shareholder debt relief; the presentation in the dilution table is misleading.

The PIK loan as a value-transfer mechanism. Issued in the March 2020 restructuring to a “syndicate of financial institutions” that were the shareholders, at 14.0% (13.5% payment-in-kind plus 0.5% cash). The balance compounded from $728M to $878M while generating $269M of interest expense across FY2024–25 alone, producing large reported losses and — because UK corporate-interest-restriction rules cap deductions — mostly valuation-allowanced tax losses rather than cash savings. A $50M partial repayment in April 2024 was funded by new third-party term debt. Then, three months before pricing, the owners forgave 85% and took $155M of the residual back in cash out of the offering. The genuine benefit to new shareholders is real and should be credited: cash interest falls by roughly $39M a year. But the structure’s history says clearly how the controlling owners have treated the capital account.

An undisclosed conflict worth flagging. J.F. Lehman acquired Forged Solutions Group in March 2026 — Sheffield-headquartered, making “flight-critical rotating aeroengine components” in nickel and titanium — as well as Wellman Dynamics (aerospace castings). JFL’s two DPC board nominees (Harman, and Stanley Deal, who sits on DPC’s compensation committee and is deemed “independent” despite being a JFLCo Operating Executive) have roles across both, and DPC’s articles waive corporate opportunities. The three facts are disclosed in three separate sections of the prospectus and connected in none; there is no risk factor addressing the overlap. This does not establish wrongdoing, and forgings are adjacent rather than identical to castings, but a sponsor building a competing aeroengine-components platform while controlling 16% of DPC and nominating its compensation-committee members is a conflict shareholders should see stated in one place.

M&A, dividends and structure. One acquisition since 2020 (Uni-Pol, 2022) with no price, multiple or return disclosed, and Uni-Pol Mexico’s $9M of goodwill now fully impaired. Ivostud has been held for sale for two years and remains unsold. The prospectus reserves the right to use proceeds for “potential strategic acquisitions” — with a rich multiple and a ~$7.3B enterprise value, issuing paper for assets is theoretically attractive, but the disclosed execution record on both acquisitions and disposals is thin. No dividend is anticipated. The Jersey incorporation brings a classified board, blank-cheque preferred stock, no pre-emption rights, no appraisal rights, a Jersey exclusive forum for fiduciary claims, and an Enforcement of Civil Liabilities section acknowledging “uncertainty as to whether the courts of Jersey would recognize and enforce judgments of U.S. courts” — a practical remedy gap that compounds the material-weakness risk. The prospectus also omits the required total for officers and directors as a group.

Insider transactions. The nine Form 4s filed 2026-06-26 report approximately $75.8M of code-P purchases, which reads as a strong alignment signal until it is decomposed: roughly 80% is recycled incentive-plan cash. Genuine personal capital was $15.45M ($7.2M of director participation in the Concurrent Private Placement plus $8.25M via the Directed Share Program), against $60.37M of MIP-funded reinvestment and 62,500 free matching shares. Director and co-chair Dirkson Charles put in $7.0M at $33.00 personally, which is meaningful. There have been zero insider sales — but that is mechanically guaranteed by the lock-up and therefore uninformative.

Lock-up and overhang. The 180-day lock-up from the 2026-06-24 prospectus date expires around 2026-12-21, releasing roughly 117.1M shares (about 80% of the company, ~$5.8B at $49.91) against a free float of only ~19–21% — over three times the float. Jefferies and Morgan Stanley may waive it at their sole discretion at any time. There is no price or earnings early-release trigger. Separately, 76,563,521 shares carry registration rights with three S-3 demands and three takedowns a year. QIA’s 2,272,727 shares are restricted but explicitly not locked up, and QIA is a weaker anchor than it appears: the placement was agreed 2026-06-22, two days before pricing, brokered by the syndicate for a 4.5% ($3.375M) placement fee, for 1.57% of the company, with no lock-up commitment.

Verdict: capital allocation is the weakest leg of this thesis. Real positives exist and should be weighted: genuine third-party deleveraging worth ~$39M a year of cash interest; J.F. Lehman committing ~$143M of its own money at $33.00 while selling nothing; no selling shareholders in the offering at all; an escalating-strike option ladder that requires ~10% annual compounding to pay; and $60M of incentive proceeds voluntarily reinvested. But they are outweighed. Forty-four per cent of net proceeds went to insiders and existing owners while 18% stayed in the business. The incentive hurdle was deleted in March 2024 and the payout re-based to the IPO price, with the CEO on his own remuneration committee. $121M of compensation cost never touched an income statement and $280M was added back out of the margin story. The $774M “shareholder contribution” is a self-forgiven related-party loan presented as invested capital. There is no returns-based metric of any kind, no say-on-pay vote, and a compensation committee chaired by a $23.9M beneficiary. And a Jersey structure leaves shareholders with little practical remedy — ahead of an ~80% share unlock on 2026-12-21. Management has been rewarded extraordinarily well for a turnaround that the disclosed record shows was substantially cyclical.

8. Changes and Headwinds — Last Two Years

The two-year window here is unusually consequential, because it contains the entire transformation of the capital structure and the company’s arrival as a public entity.

March 2024 — the incentive hurdle is removed. The management incentive plan’s thresholds, which had escalated at 13.5% quarterly, were amended so that “the 13.5% compounding was removed,” an amendment the accounts say “led to a significant increase in the charge and liability.” This is the origin of the $29M → $87M jump in incentive expense across FY2024–25 and of the eventual $280M payout. It is the single most value-relevant governance event of the period and it received no external attention.

April 2024 — a $50M PIK repayment funded by new third-party debt. A partial repayment of the shareholder loan, funded not from cash generation but from fresh borrowing, which also triggered payouts under the incentive plan.

August 2023 (just outside the window but the driver of everything inside it) — the Siemens Energy contract. Doncasters signed an agreement “worth over $1 billion USD” with Siemens Energy, described as the backbone of significant capital investment in airfoil casting plants. This mechanically explains the IGT ramp — +25.4% in FY2025, 78% of the year’s revenue growth — and makes Siemens Energy the probable 22%-of-revenue Customer A.

2024–2026 — the four strategic customer partnerships. Signed progressively from 2024: the 15-year aerospace blades-and-vanes agreement, the 7-year and 9-year IGT extensions with guaranteed volumes, and the three agreements underwriting greenfield superalloy expansion. Together “>$200M” of incremental annual revenue with up to 80% customer capital contribution, reaching full run rate in 2028 with full-year effect in 2029. Alongside them, announced plant investments: approximately $52M at Bochum (a 3,000-square-metre production hall from spring 2026 plus a series of DS furnaces on a 12–18-month installation programme), DS/SX capability at Oxford, Alabama, an end-to-end conversion at Mexicali, and a doubled shell line at Groton — the last of which took Groton production offline during 2025 and explains why aerospace grew only 9% in a strong year.

December 2025 — the corporate reorganisation and the PIK forgiveness decision. The holding company, incorporated in Jersey as “Alloy Topco Limited” in November 2019, was renamed DPC Holdings Limited. In the same month shareholders unanimously consented to reduce the $878M Shareholder PIK Loan principal by 85%.

March 2026 — the forgiveness takes effect. Effective 2026-03-19, approximately $730M was forgiven and credited to additional paid-in capital, taking total debt from $1,434M to $712M and shareholders’ equity from $(964)M to $(237)M. Also in March 2026: the India facility halted production for two weeks on LPG unavailability tied to the Middle East conflict, taking five weeks to return to full rate; and, separately, J.F. Lehman acquired Forged Solutions Group, a Sheffield-based maker of flight-critical rotating aeroengine components.

June 2026 — the IPO. A 4-for-1 reverse share split on 2026-06-05. The deal launched at roughly $700M, was marketed at $28–32, and priced above range at $33.00 on an upsized 27,858,585 shares on 2026-06-24, raising $919.3M gross, with a $68.76M insider placement and a $75M QIA placement agreed two days before pricing. Eleven Form 3s and, two days later, nine Form 4s. The stock opened at $44.00 and closed its first day at $46.88, +42%.

July 2026 — the initiation wave. The 25-day prospectus-delivery period ended 2026-07-19. On 2026-07-20, four of the six syndicate banks initiated coverage — Jefferies Buy, Morgan Stanley Equal-weight, RBC Outperform, Rothschild & Co Redburn Buy — and management set the first results date for 2026-08-11. The stock’s all-time high of $53.50 is exactly the highest published target. No Barclays or Moelis note has appeared, leaving two further initiations outstanding.

Headwinds accumulating into FY2026. Several are quantifiable and worth listing plainly. Cobalt entered 2026 at ~$56,414 a tonne, the highest since July 2022, after DRC export quotas; Indonesia cut its 2026 nickel mining quota to 260–270Mt from 379Mt — and while index-linked pass-through protects dollars, it dilutes percentage margin and lags. Q1’26 operating cash flow was negative $(7)M on a $43M inventory build attributed directly to rising input metal costs. Committed capex went from $11M to $97M in fifteen months. The corporate cost wedge widened from $2M to $5M a quarter. Turbo Wheels deteriorated to a 3.7% margin with 100% decremental drop-through. Operating leverage halved, with incremental EBITDA margin falling from 45% to 22%. $30–40M a year of stock compensation is coming and is disclosed nowhere. Three key contracts representing ~10% of revenue expire within one to two years. And ~80% of the share count unlocks around 2026-12-21.

Verdict: the changes strengthen the balance sheet and the contracted revenue base, and weaken the governance case and the quality of the reported numbers. On the business, the direction is genuinely positive: four multi-year contracted partnerships that no peer publicly matches at this scale, a doubled backlog-to-revenue relationship, real capacity being installed with customer money, and an entry into the aftermarket-rich aero airfoil category. On the capital structure, the change is transformational and permanent — from 10.2x total leverage to net cash, worth ~$39M a year. On governance and disclosure, the same period contains a hurdle deletion that enriched nine people by hundreds of millions, a related-party write-off timed three months before pricing, a $121M compensation charge routed around the income statement, three material weaknesses, and an unaddressed sponsor conflict. The company that exists today is genuinely stronger than the one that entered 2024. The question this analysis keeps returning to is not whether it improved, but what that improvement is worth at 48x.

9. Risk Analysis (Risk Matrix)

Reading the matrix. The dominant risks are not existential — they are valuation and cycle risks, and they are concentrated in the top four rows. There is no realistic path to a total loss here: the company is net cash, has no pension deficit, holds a $930M firm order book with twenty-year customer relationships, and has no going-concern qualification. The catastrophic-loss scenario is a quality escape on a safety-critical rotating or hot-section part, which is genuinely low-probability given the qualification regime but would be severe — and it is notable that DPC’s prospectus has no dedicated product-liability or quality-escape risk factor at all, while it does have one about “our use of AI.”

The realistic bad outcome is prosaic and it is a combination of rows 1, 2, 4 and 7: the cycle pauses or merely normalises, capex and working capital consume the cash the EBITDA implies, margin convergence stalls in the high teens rather than reaching the mid-twenties, and the multiple travels from ~48x toward the low twenties. Nothing in that sequence requires a disaster. Row 6 — the December lock-up — is the one dated, mechanically certain event on the calendar, and it lands on a stock whose price has been set by a fifth of its shares.

10. Valuation Discussion (Embedded Expectations)

Where the stock trades. At $49.91 on 145,151,799 shares (the prospectus basis; deal-counsel descriptions of a 32,037,372-share offering indicate the over-allotment option was exercised in full, which would give 149,330,586 shares — the 2026-08-11 10-Q cover will settle it), DPC carries a market capitalisation of roughly $7.25–7.45B. Pro-forma cash of $210M exceeds borrowings of $195M, so enterprise value is approximately $7.2–7.3B; the figures below use $7.23B, and the greenshoe question moves every multiple by about 1%.

Three of those lines carry the argument. EV/LTM adjusted EBITDA of 48.5x. Pro-forma earnings of roughly 100–140x, because interest expense collapses from $222M to $5–9M but $30–40M a year of undisclosed stock compensation arrives to replace the incentive plan. And free cash flow of approximately zero on the committed capex and a growth-consistent working-capital build.

The comparison that matters: DPC versus Howmet. DPC’s own prospectus names its two competitors as Precision Castparts (private) and Howmet, so Howmet is the only listed pure comparable — and the factor evidence agrees, since Howmet, Carpenter and ATI form the tight aerospace-industry factor cluster that DPC is entering, with the Materials co-loading on Carpenter and ATI matching DPC’s melt business.

DPC trades at a higher EV/EBITDA multiple than Howmet. It earns roughly half Howmet’s margin, at a tenth the scale, from the #4 position with 5–6% share, with GAAP losses, no free cash flow, post-tax returns at roughly its own cost of debt, three material weaknesses and 21 trading days of history. It screens cheaper on EV/Sales — 8.2x against 13.4x — but that gap is the margin gap. A business earning 16.5% EBITDA margins should trade at a lower multiple of sales than one earning 31%; the sales multiple says nothing until the margin converges.

And the anchor is stretched. On a ten-year lookback, Howmet sat at roughly the 95th percentile of its own valuation history as of mid-June 2026 (at $264.67) — 99th on price-to-sales and price-to-book — and concluded the price embedded the entire bull operating case as base case. Howmet has since risen 9.3% to $289.26, from ~42x to ~46x. DPC is being valued off a comparator that is itself at the richest point in its history.

The wider cohort makes the same point. On EV/EBITDA: TransDigm 20.1x, RTX 20.1x, Safran 23.1x, Siemens Energy 24.6x, Moog 24.9x, GE Aerospace 32.9x, ATI 33.6x, Curtiss-Wright 34.6x, Woodward 35.5x, HEICO 37.7x, Carpenter 39.7x, Howmet 46.1x, DPC ~48x, Karman 51.9x. DPC sits in the top three of its entire sector cohort — beaten only by Karman Holdings — a 2025 aerospace IPO at a record multiple that is now some 60% off its high. The company with the weakest margin structure, the lowest returns on capital and the shortest track record in the cohort carries nearly the highest multiple in it.

What the price requires — embedded expectations. Rather than build a false-precision discounted cash flow on a company with two audited years and an unvalidated IFRS-to-GAAP conversion, invert the question: what must DPC earn for today’s ~$7.23B enterprise value to be a fair price?

Read from the bottom up. Only if DPC deserves Howmet’s own record multiple does the required EBITDA sit within reach of the current run rate. Assume instead that it eventually trades where mature, profitable, cash-generative aerospace suppliers trade — 20–25x — and the company must roughly double to triple EBITDA merely to justify today’s price, before any return to the buyer.

Now size the growth actually contracted for. The four partnerships deliver “more than $200 million” of revenue at full run rate in 2029. At a 20% incremental margin — generous, given the counterparty funds up to 80% of the plant and prices on a fixed-with-index basis — that is ~$40M of incremental EBITDA on a $138M base. Add organic growth at the observed and decelerating 30–35% drop-through:

The distribution is the thesis. The base case — in which essentially everything management has promised happens on schedule over four years — produces roughly $35–42 a share against a $49.91 price. A competent, on-plan execution of the stated strategy leaves today’s buyer with a modest loss. The bull case requires simultaneous margin convergence toward Howmet’s territory and a still-premium exit multiple, and pays roughly 45–65% over four years — an 10–13% IRR, acceptable but not compelling for a levered-operating cyclical at the #4 position with a fifth of its shares floating, three material weaknesses and its first reported quarter still ahead. The bear case, which requires nothing more exotic than a cyclical pause plus multiple normalisation, costs 60–70%.

Note where the damage comes from. The bear case above does not assume collapse — revenue still grows and margins hold at today’s level. Almost the entire downside comes from the multiple compressing from ~48x toward the 14–18x a sub-$1B-revenue, 16%-margin cyclical caster commands when the market stops treating it as a secular-growth story. Limited fundamental downside, enormous multiple downside — that asymmetry is the defining feature of the setup, and it is the same structure that makes Howmet itself vulnerable, except that DPC has none of Howmet’s cushion.

What the market is actually underwriting. Consensus forward earnings imply a forward multiple somewhere between 64x and 83x depending on the source, corresponding to roughly $0.60–0.78 of EPS and $87–113M of net income. Working backwards through a 25% tax rate, ~$7M of net interest and ~$32M of D&A, that requires roughly $155–195M of adjusted EBITDA on perhaps $1.0–1.05B of revenue — an 18–19% margin, up from 16.5%, within a year or two. The sell side has therefore already underwritten a meaningful part of the margin convergence, and the stock is still on 64–83x those earnings. There is very little left to discover on the upside, and the four published initiations cluster their targets at $47–56, i.e. within 12% of spot in either direction.

Sum-of-the-parts sanity check. The consolidated 16.5% margin conceals a two-speed business, and separating it sharpens the picture. Engine Products earned $670M of revenue at a 20.3% margin — roughly $136M of EBITDA. Turbo Wheels earned 6.5% on $185M, roughly $12M of EBITDA, and at an auto-supplier 6x is worth perhaps $70–75M — about $0.50 a share, under 1% of the market capitalisation, for 22% of the revenue and 24% of the assets. Two conclusions follow. The aero/IGT franchise carries essentially the entire ~$7.3B valuation, so the multiple on the good business is even higher than the consolidated 48x — roughly 53x on Engine Products’ own EBITDA. And Turbo Wheels’ $218M of carried assets against $70–75M of market value is live impairment exposure, which the fully-impaired Uni-Pol Mexico goodwill and the two-years-unsold Ivostud both foreshadow.

Which multiples are appropriate, and which are not. EV/EBITDA and EV/Sales are the usable measures, with EV/EBITDA the primary. Price-to-earnings is meaningful only pro-forma, and even then depends on assumptions about a tax rate that has been a credit and a stock-compensation charge the company has not disclosed. Price-to-book and price-to-tangible-book are informative mainly as a measure of how little tangible capital underpins the price: 16.0x pro-forma tangible book on a foundry whose net PP&E is $221M. And the normally decisive valuation datum — the stock’s own multi-year valuation percentile — does not exist, because DPC has 21 trading days of history; the AZI valuation index returns null and FactorsToday has no coverage at all. That absence is itself a finding: there is no own-history anchor, so the cross-sectional comparison to Howmet, Carpenter and ATI and the absolute embedded-expectations arithmetic carry the entire weight.

Verdict (no recommendation, per policy). The embedded expectations require DPC to roughly double or triple EBITDA and retain a premium multiple. The business may well double EBITDA — the contracted partnership revenue, the IGT installed-base aftermarket and the aero airfoil entry are real, and the demand backdrop is genuinely exceptional. What the price does not allow for is the ordinary: a cyclical pause, a slower ramp, margin convergence that stalls at 19% instead of 24%, a customer using its 80% capital contribution as leverage on price, capex and working capital absorbing the cash the EBITDA implies, or simply a market deciding that a 5%-share #4 caster earning its cost of debt is not a 48x asset. On the evidence, the operating story is better than the reported financials suggest and the price is considerably worse than the sales multiple suggests. The distribution of outcomes from $49.91 is negatively skewed. This section takes no position; the scenarios are provided so a reader can judge the price rather than the company.

11. Variant Perception

The consensus view. DPC is the credible third source in a structurally short, three-player oligopoly supplying the two best end markets in industrials — the narrowbody aero ramp and AI-driven gas-turbine electrification — with a management team that has already doubled revenue and tripled margins, a backlog up 28% in a quarter, four customer-funded partnerships adding >$200M of revenue, and a balance sheet transformed from 10x leverage to net cash. Three of the four published initiations are constructive, targets cluster at $47–56, and the Seeking Alpha framing is “visible customer-funded growth” with “clear visibility to 2029.” The consensus is: own the scarce asset in the scarce market, and the margin gap to Howmet is the opportunity, not the problem.

The strongest bull case. (1) The margin gap is the opportunity: Engine Products already earns 20.3% and management has line of sight to peer-level economics; every 100 basis points on $1.3B of 2029 revenue is $13M of EBITDA, and the gap to Howmet is 1,300 basis points. (2) The growth is contracted, not forecast — 15-year, 9-year and 7-year agreements with guaranteed volumes in three of four, and a $930M firm order book. (3) The customer funding is genuinely extraordinary: no peer publicly matches four partnerships with up to 80% capital contribution, and it means DPC gets the revenue without the capital, which is why reported ROIC can rise even as capacity doubles. (4) The aero blades-and-vanes entry addresses the single largest quality gap — an aftermarket-rich product category on a 15-year contract, which if delivered transforms the revenue mix by 2029. (5) The demand backdrop is not a forecast either: GE Vernova is sold out through 2030, Siemens Energy to FY2028, prices have tripled, and DPC is the marginal supplier in a market where the binding constraint is casting capacity. (6) Deleveraging is worth ~$39M a year of cash interest, permanently. (7) J.F. Lehman, which knows aerospace industrials better than the marginal buyer, put ~$143M in at $33.00 and sold nothing. (8) The option ladder requires ~10% annual compounding before management gets paid. If the cycle runs another four years and margins converge, EBITDA triples and $49.91 will look cheap.

The strongest bear case. (1) Valuation: ~48x LTM EBITDA and roughly 100–140x pro-forma earnings — above Howmet’s multiple on half its margin, from the #4 position with 5–6% share, with no own-history anchor to appeal to. (2) The scarcity rent is not a moat, and the evidence is internal: the OEMs’ selling prices roughly tripled while DPC’s adjusted EBITDA margin moved 40 basis points, and Turbo Wheels — where DPC calls itself the global market leader — earns 6.5%. (3) The capital cycle: capacity contracted in 2026 arrives 2028–31 into an order book expected to peak in 2026, with Howmet, PCC, CPP, Chromalloy, the OEMs and the melters all adding, and Berkshire’s $9.8B PCC write-down and GE’s $23B Power write-down as the last two precedents. (4) Returns do not clear the cost of capital: post-tax ROIC of 5–12% against a 10.8% term loan, and negative on GAAP EBIT. (5) No free cash flow, ever — three-period cumulative −$58M, with capex tripling and working capital consuming 28% of incremental revenue. (6) The reported numbers are not yet trustworthy: three material weaknesses including an unvalidated IFRS-to-GAAP conversion, ex-works revenue recognition, a GAAP operating loss the presentation format conceals, and no operating-income line in the filing. (7) Governance: $280M to nine people after the hurdle was deleted, $121M of it never touching a P&L, no returns metric, no say-on-pay, a comp committee chaired by a beneficiary, and a sponsor building a competing forgings platform under corporate-opportunity waivers. (8) The technical setup: a 51% move on zero new information, a ~20% float, and ~80% of the shares unlocking on 2026-12-21 into the hands of distressed-credit funds whose basis is a defaulted loan.

The three to five assumptions that actually matter.

1. Does the margin converge, and how far? Everything turns on this. Drop-through has already halved from 45% to 22%. The bull needs the mid-twenties; the evidence supports the mid-teens by 2028–29.
2. Who captures the scarcity rent — DPC or its customers? The 40-basis-point answer from FY2025 is the single most decisive datapoint in this article, and it currently says the customers do.
3. What are the actual economics of the customer-funded partnerships? Not disclosed — no pricing mechanism, no return, no asset title, no accounting. If contributions are netted against PP&E, reported ROIC and margin are partly an accounting courtesy.
4. When does the capital cycle turn, and does DPC’s 2029 revenue arrive before or after it? The qualification lag says 2028–31; the order peak says 2026. Those two dates are the whole industry question.
5. Does the multiple hold? From ~48x it is the largest single driver of forward return, and there is no own-history percentile to appeal to in either direction.

What would falsify each side. Falsify the bull: two or three quarters showing Engine Products margin flat or down while capex and working capital consume the cash; evidence that customer contributions are netted against PP&E; a Boeing/Airbus rate cut or a pause in gas-turbine ordering; or the December lock-up clearing at a materially lower price. Falsify the bear: Engine Products margins marching through the mid-twenties with positive free cash flow after the incentive plan is paid and the capex step-up absorbed; the material weaknesses remediated on schedule; the aero blades-and-vanes programme pulling forward; and the multiple simply holding — which would prove the premium is earned.

Where the positioning evidence points. This is not a stock whose price contains much information. Only ~19–21% of the shares can trade — roughly $1.4–1.6B of float against a ~$7.3B market capitalisation — and the entire 51% move since pricing occurred with no 10-Q, no 8-K, no earnings call and no guidance. The four-session run into $49.91 tracked the aero-supplier complex at roughly twice its beta after GE Aerospace raised guidance. The all-time high of $53.50 is precisely the highest published sell-side target, produced by an underwriter on the first day it was permitted to publish, on the same date the stock made its lowest close. FactorsToday has no coverage at all (21 days against a 252-day minimum), so DPC has no measurable factor loading, no risk-adjusted track record and no idiosyncratic-vol estimate — none was fabricated. What can be measured is the complex DPC is entering: Howmet, Carpenter, ATI, Moog and HEICO all sit within 1.3–5.4% of all-time highs after twelve-month gains of 55–121%, on factor signatures showing positive Momentum riding negative Value and negative-or-absent Quality — the crowded end of the trade, not the quality-compounder end. The genuinely positive-Quality aerospace names, TransDigm and HEICO, are the ones flat to down over twelve months. The Momentum factor itself has stalled in the last month (z −0.70) while Value has run hot (z +1.40).

A price set by a fifth of the shares is weak evidence about value in either direction, and that cuts against both sides: the bull cannot claim the market is validating the story, and the bear cannot claim the market is signalling a problem. The informative facts are structural, not directional — the float mechanics, the 2026-12-21 supply date on over three times the float, and the observation that the whole comparable complex is priced at cycle-high multiples on volume and cycle exposure rather than on quality.

Where I come out (see the opinion block at the top). The bull is right about the business trajectory and wrong about the price; the bear is right about the price and somewhat too harsh on the business, which is genuinely improving and has real contracted growth arriving in 2028–29. The synthesis — a good hard business, a weak relative position, an excellent cycle that is visibly late, and a multiple above the best operator in the industry — is an AVOID at $49.91 rather than a short, with the accumulation zone at or below the IPO price.

12. Fact vs. Interpretation Table

13. Open Questions

1. How are the customer capital contributions accounted for, and how large are they? Not one dollar appears in the statements. Deferred revenue as PP&E offset versus contract liability determines whether reported margin and ROIC are economically real. First-order.
2. What are the commercial terms of the four partnerships? No pricing mechanism, no return on the funded asset, no asset title, no clawback, no exclusivity terms disclosed. Without these the bond-versus-lien question cannot be settled.
3. What is stock compensation going to be? No grant-date fair value, no valuation assumptions, no expected expense disclosed anywhere for 7.5M options. Our $30–40M/yr estimate is unanchored.
4. What was revenue and EBITDA in FY2019 on the current perimeter? The single most important missing number — it would establish where FY2025’s 16.5% sits relative to the prior cyclical peak. Not disclosed; the two-year EGC presentation prevents it.
5. Who are Customer A (22%) and Customer B (16%)? Siemens Energy is the strong inference for A; B is unidentified.
6. Which three contracts (~10% of revenue) expire within one to two years, and on what terms will they renew?
7. Was the greenshoe exercised, and what is the true share count? Deal counsel implies full exercise; every data feed still shows 145.15M or fewer. The 2026-08-11 10-Q cover settles it.
8. Does the Term Loan’s maturity spring to December 2027? The PIK is being repaid rather than extended to July 2030.
9. What are the debt covenants and the headroom? Levels and tests are not disclosed.
10. Why does intangible amortisation step from $8M (2027–28) to $23M (2029–31)? No explanation, no FY2026 figure.
11. What is the environmental and occupational-disease exposure? No reserve, no dollar figure, and zero mentions of asbestos, silica, beryllium or hexavalent chromium across 836KB — unverifiable rather than clean.
12. Was there fresh-start or pushdown accounting in the March 2020 restructuring? Determines whether gross PP&E reflects anything like replacement cost.
13. Who are the four unnamed MIP participants sharing ~$76.9M?
14. Did the $774M PIK release consume the $219M of UK non-trading losses? Material to the cash tax rate.
15. Will the material weaknesses be remediated before the FY2026 10-K, and what will the first 404(a) management report say?
16. What happens at the lock-up expiry? Whether the credit-fund holders sell, and at what price, is the largest single mechanical uncertainty.
17. Is DPC eligible for major index inclusion? A float-adjusted add would be another purely mechanical demand source.
18. What explains the unexplained 2,705,514-share print on 2026-07-16?

14. What Must Be True

For the bull case to work — and the falsification test for each.

1. Margin must converge toward the mid-twenties, not stall in the high teens. Engine Products must carry 20.3% toward 26–30% while the corporate wedge stops widening and Turbo Wheels stops deteriorating. Falsification test: if the FY2026 and FY2027 reported Engine Products margins expand by less than ~100bp a year, or if consolidated adjusted EBITDA margin fails to exceed 18% by FY2027, the convergence thesis is broken and the bear’s mid-teens destination is confirmed. Watch the 2026-08-11 and subsequent prints; drop-through below 30% is the tell.
2. DPC must start capturing scarcity rent rather than passing it through. The 40-basis-point response to a tripling in OEM prices must reverse in the LTA and spot repricing rounds now under way. Falsification test: if FY2026 adjusted EBITDA margin fails to beat 16.9% while cobalt and nickel remain elevated and gas-turbine prices keep rising, DPC is a price-taker and the moat claim fails on its own evidence.
3. Free cash flow must appear. The company must fund $58–97M of capex and a 28%-of-incremental-revenue working-capital build and still generate cash. Falsification test: if FY2026 free cash flow is negative, or if the cash conversion cycle exceeds 120 days, the business has demonstrated across four consecutive periods that it cannot convert EBITDA into owner cash — which at 48x is disqualifying.
4. The customer-funded model must prove to be a bond, not a lien. The 2029 revenue must arrive at margins at or above the current corporate average. Falsification test: disclosure that contributions are netted against PP&E, or partnership revenue arriving at below-average margins, confirms DPC is renting its operating capability at cost-plus while the customer keeps the rent.
5. The reported numbers must prove trustworthy. The three material weaknesses must be remediated without restatement. Falsification test: any restatement, any additional material weakness identified at the first 404(a) assessment, or an auditor change would break the entire evidentiary basis of this analysis.
6. The cycle must run past 2029. DPC’s contracted revenue arrives in the same window as the industry’s qualified capacity additions. Falsification test: a Boeing or Airbus rate cut, a quarter of declining gas-turbine orders at GE Vernova or Siemens Energy, or gas-turbine pricing rolling over, would mark the peak ahead of DPC’s ramp.

For the bear case to work — and the falsification test for each.

1. The multiple must compress. ~48x must travel toward 20–25x as the market re-classifies DPC from secular-growth story to subscale cyclical caster. Falsification test: if DPC holds above 40x EV/EBITDA through two reported quarters and a lock-up expiry, the market has decided the premium is structural and the bear’s central mechanism has failed.
2. The capital cycle must turn on schedule. The 2028–31 capacity additions must arrive into decelerating demand. Falsification test: gas-turbine backlogs extending beyond 2032 with pricing still rising into 2028, or capacity additions slipping materially, would extend the cycle past DPC’s ramp and vindicate the bulls.
3. The lock-up must produce real selling. ~117M shares held by distressed-credit funds with a defaulted-loan basis must actually come to market. Falsification test: if the December 2026 expiry passes with minimal volume and no price impact — or if holders convert to long-term positions or sell in orderly marketed blocks at or above the prevailing price — the overhang argument dissolves.
4. Governance must actually cost shareholders money. The $280M and the incentive design must translate into value destruction rather than merely being distasteful. Falsification test: if the escalating-strike option ladder does its job — management compounding the stock above $48.31 while capital is allocated to returns-accretive capacity — then the governance criticism, while fair on process, will have been immaterial to outcomes. J.F. Lehman’s $143M at $33.00 is the bull’s best evidence here.
5. Returns must stay below the cost of capital. ROIC must fail to move durably above ~12%. Falsification test: two or three quarters at the Q1’26-annualised level (14.6% net / 11.6% gross invested capital) or better, on a larger asset base and after stock compensation, would establish that DPC earns its cost of capital — at which point the argument becomes purely about price rather than about business quality.

The single most informative upcoming event is 2026-08-11, DPC’s first reported quarter as a public company: the first external check on the Q1’26 margin, the first look at cash flow and capex under public scrutiny, the first management guidance ever given, the definitive share count on the 10-Q cover, and the first opportunity to see whether stock compensation and the customer-funded capex accounting are disclosed. The second is ~2026-12-21, the lock-up expiry. Between them, essentially every open question in Section 13 that can be resolved by disclosure will be resolved.

15. Source Appendix

This article rests on an unusually narrow primary-source base, and that constraint should be stated plainly rather than disguised by a long citation list. DPC Holdings Limited has filed no 10-K, no 10-Q and no 8-K, and has never held an earnings call. Its entire SEC corpus is 25 filings across 101 days: a DRS (2026-04-15), an S-1 (2026-05-26) and three S-1/As (2026-06-15, and two on 2026-06-22), an FWP, an 8-A12B and CERT (2026-06-24), eleven Form 3s (2026-06-24), an S-1MEF and EFFECT (2026-06-24), an S-8 (2026-06-25), the 424B4 (2026-06-26) and nine Form 4s (2026-06-26). There have been no filings since 2026-06-26.

The 424B4 prospectus dated 2026-06-24 is therefore the single primary source for every financial figure in this article — 836,000 characters containing two audited fiscal years (FY2024, FY2025, audited by KPMG LLP, Birmingham, auditor since 2021, report dual-dated 2026-04-14 and 2026-06-15) plus unaudited Q1’26 and Q1’25 interim statements. Because the issuer discloses three material weaknesses in internal control — including an inability to verify that its IFRS-to-US-GAAP conversion adjustments “are complete and accurate” — and because no Section 404(b) attestation is required for up to two more years, every figure derived from it is provisional.

Everything else is either third-party data used for cross-sectional orientation (AZI Trading price history, FactorsToday factor data, the ROIC.ai MCP, yfinance), public secondary sources used to reconstruct the pre-IPO corporate history and the industry backdrop, or public filings of comparable companies used for benchmarking. Full citations, with URLs and access dates, and explicit flags on the handful of sources reachable only through a search index, appear in Appendix B — Source Appendix.

APPENDIX A — Standard Diligence Questionnaire

DPC Holdings Limited (NYSE: DPC) · Report date 2026-07-25 · Price referenced $49.91 Supplemental to the analysis above. Labels: [F] Fact · [I] Interpretation · [A] Assumption.

General

What thoughtful questions have other investors asked about this company?

DPC has been public for 21 trading sessions and has never reported a quarter, so there is no earnings-call Q&A, no investor-day transcript and no accumulated body of buy-side questioning. The four published sell-side initiations (2026-07-20) and two Seeking Alpha pieces are the entire external analytical record. [F]

From that thin record plus this analysis, the questions that matter most — and which an investor should put to management on the 2026-08-11 call — are:

1. How are customer capital contributions accounted for, and how much has been contributed to date? Not one dollar appears anywhere in the financial statements. If netted against PP&E, reported margin and ROIC are flattered. [F/I]
2. What return does DPC earn on the customer-funded assets, and who holds title? No pricing mechanism, return, clawback or title arrangement is disclosed. [F]
3. What is stock-based compensation expected to be? No grant-date fair value or valuation assumptions are disclosed for 7.5M options. [F]
4. Why is there no operating-income line in the income statement, and can management confirm GAAP operating income was $(5)M in FY2025 against +$31M in FY2024? [F]
5. What was revenue and adjusted EBITDA in FY2019 on the current perimeter? The only benchmark that would show where 16.5% sits versus the prior cyclical peak. [F]
6. Does the Term Loan maturity spring to 5 December 2027 now that the PIK loan is being repaid rather than extended to July 2030? [F]
7. Why was the 13.5% MIP hurdle escalator removed in March 2024, who approved it, and what would the payout have been had it remained? [F]
8. When will the three material weaknesses be remediated? [F]
9. What are the covenant levels and headroom on the remaining debt? Not disclosed. [F]
10. Why does intangible amortisation step from $8M (2027–28) to $23M (2029–31)? [F]

The bull-side question one should expect other investors to press hardest: if the margin gap to Howmet is 1,300bp and Engine Products already earns 20.3%, why should the market not capitalise the convergence today? The answer developed above is that drop-through has already halved from 45% to 22%, the corporate cost wedge is widening, and the destination the evidence supports is the mid-teens by 2028–29, not the mid-twenties. [I]

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? A cyclical high, and an unusually clear one. Heavy-frame gas-turbine prices have risen 195–300% in three years; 110 GW of orders sit against 60–70 GW of industry capacity; GE Vernova is sold out through 2030 with over half of 2031 contracted; Wood Mackenzie expects gas-turbine orders to peak in 2026. Commercial aero is mid-cycle on original-equipment build and early in the higher-margin spares cycle. DPC’s FY2025 margin of 16.5% and Q1’26’s 16.9% are therefore struck at or near a peak in its two main end markets simultaneously. [F for the market data; I for “cyclical high”] Note the countervailing point: DPC’s absolute margin is far below what the industry’s best operator earns, so “cyclical high” describes the environment, not a peak level of achievable profitability.

Driven by the external environment or internal actions? Both, and the split is decomposable only imperfectly because the filing provides no price/volume/mix/FX bridge. [F] My read: the FY2024→FY2025 350bp margin gain is mostly operating leverage on IGT volume (+25.4%, 78% of the year’s revenue growth) plus shortage-enabled repricing, with modest genuine productivity, an unquantified FX tailwind of ~17–22% of revenue growth, and at least $6M of allocation noise (Turbo Wheels’ “improvement” was a corporate recharge credit falling from $8M to $2M). The five-year doubling from ~$365M is mostly recovery from a distressed 2020 base — COVID plus a scheme of arrangement. Genuine internal action is real but secondary: capital retained through the trough while PCC cut 30% of its workforce, $170M of capex, a doubled Groton shell line, and the repricing of every key LTA. [I]

How stable are revenues? Moderately stable in composition, cyclically unstable in level. ~70% of revenue sits under LTAs typically five years or longer with index-linked pass-through; ~40% of castings revenue is aftermarket on a 20-year consumption cycle; average customer tenure exceeds 20 years; the firm order book is $930M covering more than twelve months. [F] Against that: the LTAs guarantee minimum market share rather than volume (volume commitments exist in only three of four partnerships), certain LTAs are terminable for convenience, certain have no minimum purchase commitments, and three key contracts representing ~10% of revenue expire within one to two years. [F] The demonstrated peak-to-trough revenue experience for the group was roughly 55%. [F/I — perimeter changed]

Outlook for products/services? Strong for Engine Products, weak for Turbo Wheels. Aero structural castings ride the 737 MAX ramp (38 → 47/month cleared, targeting 53 by end-2026) and the A320 family (to 75/month by mid-2027), with LEAP tracking above 2,000 engines in 2026. IGT airfoils ride record heavy-frame demand plus a growing installed-base aftermarket. The new 15-year aero blades-and-vanes programme is the most valuable product addition, taking DPC into an aftermarket-rich category from 2029. Turbo Wheels declined 2.0% in FY2025 and is flat in Q1’26 at a 3.7% margin, structurally exposed to ICE displacement. [F for the data; I for the outlook]

How big will this market be — growing, shrinking, domestic or international? My bottom-up estimate of the served market is $7–9B for merchant nickel/cobalt aero and IGT castings plus $3–4.5B for premium melt, ~$10–13B in total, ±30% — excluding OEM in-house casting, which is unquantifiable. DPC holds roughly 5–6%, against #1 and #2’s combined 55–70%. [I/A — DPC discloses no market-share or TAM figure anywhere in 836KB of prospectus, which is itself a finding.] Growing in both aero and IGT on a multi-year view; shrinking in transportation. Thoroughly international: 64% of FY2025 revenue was non-US, with plants in the UK, Germany, the US, Mexico, India and China.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More, at the margin, and deliberately so. The OEMs are actively qualifying third and fourth sources — DPC’s own prospectus says the partnerships provide “additional capacity and competition within these supply-constrained markets” and that “customers are actively diversifying their supply chains.” Howmet’s capex is guided to $470–500M with ~70% into Engine Products; PCC is expanding Ravenswood explicitly to enable “capacity-based pricing” and made its first post-Berkshire acquisition (Morvern Group, Derby and Worcester) in March 2026; CPP is building in Euclid, Ohio; Chromalloy has entered the aftermarket airfoil pool with its own FAA PMA; the OEMs are adding in-house capacity; and the premium melters are adding furnaces qualifying to ~2030. [F]

How profitable is the business (ROIC, ROE)? Post-tax ROIC on adjusted EBIT was 12.4% on net invested capital ($641M) and 9.8% on gross invested capital ($810M) in FY2025. On GAAP EBIT it was negative. Levelizing the $280M incentive plan over its 6.29-year life gives 6.6% and 5.2%. Against a disclosed 10.8% term-loan rate and a 13.0% lease discount rate, returns are at or below the cost of capital. [F/computed] Peer ladder on comparable bases: Howmet ~21% > Carpenter 17.2% > ATI 13.7% > DPC ~10–12%. [F/I] ROE is meaningless pre-IPO (equity was negative at all three balance-sheet dates); pro-forma post-IPO ROE is ~9–12% on $625M of equity. [F] The trend is improving on every definition, and Q1’26 annualised (14.6% net / 11.6% gross) is the first reading that plausibly clears a hurdle rate — one annualised quarter is not a track record. [I]

How profitable is the industry — how many competitors, what barriers to entry? The industry’s best operator earns 33.3% segment EBITDA margins (Howmet Engine Products, a record 36.6% in Q1’26) and ~21% ROIC, so the industry is genuinely profitable at the top. Four globally scaled casters (PCC, Howmet, CPP, DPC), HHI roughly 1,900–2,400 — concentrated but not a duopoly. Barriers are real: NADCAP/AS9100 accreditation takes over a year plus substantial capital; OEM part-level process approval “often tak[es] years” and is specified to individual process parameters; critical quality roles require 6–12 months of operator training; melt furnaces qualify grade-by-grade over years; and specialised capital equipment has long lead times. Net, capacity announced in 2026 is not sellable qualified capacity until roughly 2028–2031. [F] Two things those barriers are not: they are not barriers DPC controls (it holds no FAA production approvals of its own and casts to customer-owned designs and tooling), and they do not impede expansion by already-qualified incumbents — which is exactly what is happening. [I]

Can the business be easily understood? Yes, at the level of what it does — it melts superalloy and casts hot-section engine parts to customer print. No, at the level of what it earns: the reported P&L has no operating-income line, adjusted EBITDA excludes a $280M cash cost, $121M of that cost bypasses the income statement entirely, the tax line was a credit driven by a valuation-allowance release, the customer-funded capex is quantified nowhere, stock compensation is undisclosed, and the whole set rests on an IFRS-to-US-GAAP conversion the company says it lacks the expertise to validate. [F] This is a simple business with genuinely difficult accounts. [I]

Can it be undermined by foreign low-cost labour? Not materially in Engine Products. The binding constraints are qualification, metallurgical know-how and OEM approval rather than labour cost, and safety-critical hot-section content for Western engine programmes is not readily re-sourced to a low-cost entrant — DPC itself takes 6–12 months to accredit a critical quality operator. [F/I] Turbo Wheels is a different matter: it already operates in China, India and Mexico, buys its alloy externally “primarily from suppliers in China,” and earns 3.7%. That is a labour-and-commodity-cost business and it shows. [F]

Do brands matter? No. This is a make-to-print component business selling to nine sophisticated industrial buyers; the “Doncasters” name carries reputational weight in the supply base (an RTX Platinum Award, 20-year relationships) but confers no pricing power. The relevant intangible is qualification status on specific part numbers, and that asset is held jointly with — and revocable by — the customer. DPC’s only material patent estate is 34 Chinese patents on turbocharger wheels. [F/I]

What is the nature of competition? Competition for share of work on specific part numbers within a qualified supply base, decided by an OEM on capacity availability, on-time delivery, yield, quality record and price, and increasingly by the OEM’s strategic desire to hold two or more sources. It is not price competition in a commodity sense, nor differentiated-product competition; it is allocation competition inside an approved vendor list. In shortage the qualified supplier holds the whip; in surplus the OEM does. [I] The empirical test: the OEMs’ output prices rose 195–300% while DPC’s adjusted EBITDA margin moved 40 basis points. [F]

Customers’ switching costs? High per part number, low at the relationship level, and — critically — paid by the customer, not by DPC. Re-sourcing a flying hot-section part requires OEM process re-approval taking years and carries regulatory and schedule risk. But the customer owns the design and the tooling, holds the FAA production approval, bears the qualification cost, and can dual-source, re-source or in-source (all three are named risk factors). And with capacity “allocated to that customer only” under the funded partnerships, the asset-specificity runs the wrong way: the party holding the single-customer-specific asset is the weaker one at renewal. [F for the disclosures; I for the conclusion] The clean internal falsification is Turbo Wheels, where identical technology and claimed global market leadership produce a 6.5% margin because the end market is not short and the buyers are Tier-1 automotive.

Financial Condition & Balance Sheet

Assets not fully recognised on the balance sheet? Probably yes, in two forms. First, replacement cost: gross PP&E is $337M, or 0.40x revenue — roughly half Howmet’s ratio — on an asset base carried at values set by acquisition accounting around the 2020 restructuring. For 14 plants including three vacuum-induction melt shops, this likely understates replacement cost. [I] Second, the qualification asset: multi-year OEM part approvals across a large part portfolio have real economic value and appear nowhere on the balance sheet — though as noted, that asset is co-owned with the customer. [I] Also unrecognised: $219M of UK non-trading losses, $105M of German losses and $184M of interest carryforwards, of which only part is recognised as a deferred tax asset ($44M net after a $40M valuation allowance). [F]

Off-balance-sheet liabilities? None disclosed, and this is a genuinely clean answer. The company states: “We do not currently engage in off-balance sheet financing arrangements. In addition, we do not have any interest in entities referred to as variable interest entities.” Explicit searches for factoring, receivables purchase, supply-chain finance, supplier finance, non-recourse and securitisation return zero hits across 836KB. [F] Lease obligations are small ($16M at FY2025, $18M at Q1’26) and on balance sheet. Two letters of credit sit against the ABL relating to UK energy hedging and intra-group financing, described as immaterial. [F] The genuine unquantified exposures are: environmental and occupational-disease legacy for a ~248-year-old foundry group with formerly-owned sites (CERCLA and the UK Contaminated Land Regime, potentially joint and several) — no reserve and no dollar figure disclosed, and zero mentions of asbestos, silica, beryllium or hexavalent chromium anywhere — and unquantified “Legacy Provisions” for former-employee claims recognised in acquisition accounting. [F] That is unverifiable rather than clean.

How conservative is the accounting? Not conservative, and self-declared unreliable in three respects. Three material weaknesses as of 2025-12-31, identified by both the company and KPMG: no formal control framework or general IT controls; “lack of resource and expertise necessary to apply U.S. GAAP conversions from IFRS, including whether the list of adjustments are complete and accurate”; and no segregation of duties over journal-entry review and approval across components. Remediation explicitly names revenue recognition at the manufacturing sites. [F] Specific aggressive or opaque choices: revenue recognised largely ex-works (“control … transfers to the customer when the products are made available for collection at the Group’s premises”), so goods need never leave the plant; the scrap-return refund liability roughly doubled ($8M → $17M); $62M of gross valuation-allowance churn against a $(6)M net rate-reconciliation line in the year that set the offering price; a Level-3 valuation on the MIP liability discounted at 25%; no operating-income line; the $87M MIP charge added back to adjusted EBITDA while $121M more bypasses the P&L entirely; and an unexplained amortisation step from $8M to $23M in 2029. [F] Conservative choices that deserve credit: FX gains are removed from adjusted EBITDA; disposal-group impairments and reversals are treated symmetrically; the PIK forgiveness was correctly credited to APIC rather than taken as a P&L gain; and there is no going-concern qualification. [F]

How CapEx-hungry is the business? Historically much less than one would expect, and about to become much more. FY2025 capex was $31M — 3.7% of revenue, and down 14% from $36M — against $32M of D&A. Cumulative capex since 2020 is “more than $170 million,” roughly $28M a year against ~$22–23M of annual depreciation, so six years produced only $5–6M a year of net asset growth: the plant has been maintained, not expanded. [F] Note that capex/depreciation of 1.35x (FY25) and 1.64x (FY24), with the base only 34.4% depreciated, means the reported EBITDA is not being harvested from a starved plant — I tested that hypothesis and it is not supported. [F] Forward, committed-but-unspent capex went $11M → $53M → $97M in fifteen months, guidance is ~$58M over the next twelve months, and there is a greenfield superalloy plant to build. Capex is therefore moving from 3.7% of revenue to 7–12%. The offsetting claim — that customers fund up to 80% — is quantified nowhere in the statements. [F]

Capital Allocation & Management

How much FCF does the business generate, how does management use it, what is the philosophy? It does not generate free cash flow. FY2024 $(52)M, FY2025 +$11M, Q1’26 $(17)M — cumulatively negative $58M. Operating cash flow was negative in two of three periods, and FY2025’s +$42M falls to roughly $3M once the 20-day trade-payables stretch is removed; the $92M increase in the unpaid MIP accrual also flattered it. Operating cash flow as a percentage of adjusted EBITDA was 30%, (18)% and (18)%. [F/computed] Pro-forma, on guided capex free cash flow would be ~$60–70M (a 0.8–1.0% yield); on the $97M of committed capex plus a working-capital build at the observed 28.1% of incremental revenue, pro-forma free cash flow is approximately zero. [I/computed] The stated philosophy is reinvestment for growth with no dividend. Historically, capital has gone to maintenance capex, debt service on a 14% related-party PIK loan, and — in 2026 — $280M to nine people. [F]

Significant acquisitions recently? One since 2020: Uni-Pol (2022), with no price, multiple, contribution or return disclosed, and Uni-Pol Mexico’s $9M of goodwill now fully impaired — at the very site being converted into an aerospace hub. On the disposal side, Ivostud (a German stud-welding unit, 169 employees) has been held for sale since 2024, written down $(9)M in FY2024 and written back +$5M in FY2025, and remains unsold after two years on a letter of intent. [F] The prospectus reserves proceeds for “potential strategic acquisitions” with no commitments. Given a rich multiple, issuing paper for assets is theoretically attractive; the disclosed execution record on both acquisitions and disposals is thin, and empire-building at a sector peak is a real risk. [I]

Buying back shares? No. No buyback authorisation exists and none would be appropriate — the company is one month public, has an ~80% lock-up expiry ahead, needs to fund a capex step-up, and generates no free cash flow. [F/I]

Issuing large amounts of new shares to insiders? Yes. The equity plan reserves 13,812,500 shares — 9.52% of shares outstanding — including 7,499,995 options granted at the IPO, refreshed at 1.25M a year, with the annual cap disappearing entirely once the stock trades above $66 (twice the IPO price) after June 2027. [F] The important corrective, which cuts in shareholders’ favour: the 6,249,995 IPO Grants are not struck at $33.00 but in five tranches at $33.00 / $36.30 / $39.93 / $43.92 / $48.31, cliff-vesting in years one to five — an average strike of $40.29 requiring roughly 10% annual compounding, with intrinsic value at $49.91 of about $81M rather than the ~$127M a flat-$33 assumption implies. This ladder is genuinely well designed and is the single best element of the compensation architecture. [F/I] Separately, ~1,250,000 immediately-vested options with no performance condition were granted as consideration for agreeing to the MIP amendment (~$21M intrinsic). [F] No grant-date fair value, valuation assumptions or expected expense are disclosed anywhere; I estimate $30–40M a year of coming charges. [A]

Compensation policy of directors/management? The weakest area. $280M of IPO proceeds — 28.3% of net proceeds, 2.03x FY2025 adjusted EBITDA, 33.5% of revenue, $1.93 a share — pays nine people, triggered by the act of listing: Quinn $85.8M, Egan $38.1M, Mays $25.4M, and two non-executive directors, Charles $29.9M and Sanders $23.9M ($203.1M or 72.5% to five people; four unnamed share ~$76.9M). [F] The hurdle was lowered twice, both times increasing the payout: the 13.5% quarterly escalator was removed in March 2024 — the note states this “led to a significant increase in the charge and liability” — and the plan was re-based to the $33 IPO price at listing. The CEO sat on the committee that set his own award: “the members of our then Nominating and Remuneration Committee consisted of Mr. Charles, Mr. Sanders and Mr. Quinn. Mr. Quinn was and is our CEO” — all three were beneficiaries. The go-forward committee is chaired by Sanders, a $23.9M beneficiary, and includes Stanley Deal, deemed “independent” while being a J.F. Lehman Operating Executive. [F] The FY2025 Summary Compensation Table shows CEO total pay of $1,294,155 with zero in the non-equity-incentive column — in the year of the $87M charge, three months before an $85.8M payday. [F] The go-forward annual bonus is 100% discretionary, with no ROIC, no EBITDA, no revenue and no relative-TSR metric and no performance share units anywhere. As an emerging growth company there is no say-on-pay vote. [F] For a company whose returns sit at roughly its cost of debt and whose value depends on whether customer-funded capacity earns an adequate return, the absence of any returns-based measure is a material governance failure. [I]

Motivations of management? Mixed, and worth stating on both sides. Pre-IPO, insiders owned just 0.44% of the company — four of nine directors owned zero shares and the CEO owned 27,729 shares — so alignment has been created as cash out, not equity in. The $75.8M of code-P “purchases” on the Form 4s is roughly 80% recycled MIP cash: genuine personal capital was $15.45M ($7.2M of director participation in the Concurrent Private Placement plus $8.25M via the Directed Share Program) against $60.37M of MIP-funded reinvestment plus 62,500 free matching shares. Zero insider sales, but that is mechanically guaranteed by the lock-up and therefore uninformative. [F] On the other side: reinvesting $60M of after-tax proceeds is a choice, not an obligation; director and co-chair Dirkson Charles put $7.0M of his own money in at $33.00; and the escalating-strike option ladder means management now needs the stock above $48.31 to collect the top tranche. Going forward, management’s incentives are reasonably well aligned — it is the historical transfer, not the prospective arrangement, that is objectionable. [I] Credibility note: CEO Quinn was a PCC Group Vice President 2005–2014 and COO Mays spent 1990–2006 at Howmet; DPC’s operating leadership was hired out of the two companies it competes with, which is the most credible mechanism behind the turnaround — and a key-person risk rather than an institutional moat. [F/I]

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer? None of the three. DPC is a direct listing of Jersey-incorporated ordinary shares on the NYSE (ISIN JE00BPK4NV88); the ROIC.ai profile confirms is_adr = false. It is UK tax-resident and Jersey-incorporated, taxed as a corporation — no K-1, no MLP pass-through. [F] But the practical consequences of foreign incorporation resemble an ADR’s in one important respect: a classified board, blank-cheque preferred stock, no pre-emption rights, no appraisal rights, Jersey as exclusive forum for fiduciary claims, and an Enforcement of Civil Liabilities section acknowledging “uncertainty as to whether the courts of Jersey would recognize and enforce judgments of U.S. courts … you may not be able to collect any damages.” [F] Combined with three material weaknesses, the remedy gap matters more than it usually would. [I] DPC files as a domestic issuer (S-1, Form 3/4s), so insider reporting exists — unlike a foreign private issuer.

Dividend policy? None. “We currently do not anticipate paying any cash dividends after this offering and for the foreseeable future,” subject to debt-instrument restrictions and the Jersey requirement that dividends be paid only out of surplus or current net profits. [F] Appropriate given a capex step-up from $31M to $58–97M+ and no free cash flow — but also a signal about cash generation rather than purely a growth choice. [I]

How profitable is the business? Gross margin 23.1% (FY2025), rising to 24.1% in Q1’26. Adjusted EBITDA margin 16.5%, or 10.6% levelizing the incentive plan. Adjusted EBIT margin ~12.7%. GAAP operating margin was −0.6% in FY2025 against +4.2% in FY2024. GAAP net margin −20.7%. Engine Products, the good half, earned 20.3%; Turbo Wheels 6.5%. Post-tax ROIC ~10–12% on adjusted EBIT, negative on GAAP EBIT. [F/computed] Against Howmet’s 24.8% operating margin, 31% EBITDA margin and ~21% ROIC, DPC captures roughly 55–60% of best-in-class unit economics. [I]

Is net income diverging from cash from operations? Yes, in both directions and for different reasons — and the divergence is the most informative thing in the accounts. FY2025: net loss $(173)M against operating cash flow of +$42M, a $215M gap bridged by $150M of non-cash interest on the PIK loan, $87M of accrued-but-unpaid MIP, $32M of D&A and $(40)M of deferred tax, less a working-capital drag. So reported losses overstate the economic problem. [F] But the cash flow is not what it appears either: strip the 20-day payables stretch and FY2025 operating cash flow falls to roughly $3M; the $92M MIP accrual increase was also unpaid and is now being settled in cash out of IPO proceeds. Q1’26 shows the reverse — a $(47)M net loss and $(7)M of operating cash flow, with a $43M metal-driven inventory build. [F] The honest summary: neither the GAAP loss nor the reported operating cash flow is a fair measure of earning power. Adjusted EBITDA of ~$138–149M is the fairest gross measure; free cash flow of approximately zero is the fairest net one. [I]

Risks & Downside

What factors would cause the stock to decline? In descending order of probability-weighted impact: (1) multiple compression from ~48x toward the 20–25x that mature aerospace suppliers command, which alone accounts for most of the downside in the bear case; (2) the 2026-12-21 lock-up expiry releasing ~117M shares (~80% of the company, ~$5.8B) against a ~19–21% float, held largely by distressed-credit funds whose basis is a defaulted 2020 loan; (3) the 2026-08-11 first print disappointing on margin, cash flow or capex, or revealing that customer contributions are netted against PP&E; (4) capital-cycle normalisation as 2028–31 capacity lands into an order book expected to peak in 2026; (5) margin convergence stalling, with drop-through already halved from 45% to 22%; (6) an accounting restatement or a further material weakness; (7) an aero build-rate cut or a pause in gas-turbine ordering; (8) cobalt or nickel spikes diluting percentage margin through the lagged pass-through; (9) a single-site disruption — there have been three realised events in five years (a 2021 plant fire, a 2025 third-party X-ray outage, the March 2026 India LPG shutdown); (10) loss or adverse renewal of one of the three contracts (~10% of revenue) expiring within one to two years. [I, on disclosed facts]

Risk of a catastrophic loss? The genuine catastrophic scenario is a quality escape on a safety-critical rotating or hot-section part — an in-service failure traced to a DPC casting, which would bring liability, programme removal, NADCAP/AS9100 suspension and reputational loss across a concentrated customer base simultaneously. Probability is low given the qualification regime, HIP, X-ray and NDT controls and a clean disclosed record. It is striking that the prospectus contains no dedicated product-liability or quality-escape risk factor — the word “escape” appears zero times — while it does carry one about “our use of AI.” [F] Secondary catastrophic paths: total loss of a single sole-source facility (shell-line replacement takes up to 18 months, and 42% of revenue flows through two IGT plants), or a material accounting fraud, which the three disclosed material weaknesses make less remote than usual. [F/I]

Chance of a total loss? Very low. Post-IPO the company is net cash (~$210M against $195M of borrowings), pension liabilities total $26M with the UK schemes fully bought out, there is no going-concern qualification, the order book is $930M, and customer relationships average over twenty years. Note the company has been through a debt-for-equity restructuring once, in March 2020, when £900M was written off — so the tail is not theoretical for this asset. But that occurred at 10x leverage in the worst aerospace downturn on record; the post-IPO structure is not remotely comparable. This is a valuation risk, not a solvency risk. [F/I]

Recent News & Events

Has the business environment changed recently? Yes, decisively and favourably on demand, and unfavourably on supply-side discipline. GE Vernova’s gas-turbine backlog moved from 80 GW at end-2025 to 116 GW in Q2 2026, sold out through 2030 with over half of 2031 contracted; Siemens Energy holds 60 GW firm plus 27 GW reserved; prices are up 195–300% in three years. GE Aerospace’s July 2026 print (revenue +24%, EPS $2.02 against $1.86 expected, FY26 guidance raised, Commercial Engine Services orders +18%) re-rated the whole engine supply chain and is the identifiable driver of DPC’s own 10.8% four-session move into 2026-07-24. [F] Simultaneously, the supply response has become unmistakable — Howmet’s capex to $470–500M, PCC’s Ravenswood expansion and Morvern acquisition, CPP’s Euclid facility, Chromalloy’s FAA PMA, OEM in-house additions, and new melt capacity at Carpenter and ATI. [F] DPC’s own backlog rose $725M → $930M in one quarter. [F]

Significant acquisitions? None by DPC. But two adjacent transactions matter: PCC acquired Morvern Group (wax patterns, ceramic core assemblies, tooling; Derby and Worcester, ~£22M revenue) on 2026-03-01 — its first acquisition since the Berkshire takeover, in DPC’s home town. And J.F. Lehman, DPC’s largest shareholder at 16.00%, acquired Forged Solutions Group in March 2026 — Sheffield-based, making “flight-critical rotating aeroengine components” in nickel and titanium — plus Wellman Dynamics (aerospace castings). JFL nominates DPC directors including a compensation-committee member, and DPC’s articles waive corporate opportunities. The three facts are disclosed in three separate prospectus sections, connected in none, with no risk factor. [F]

Change in accounting policies? No policy change is disclosed, but three structural accounting events dominate the period: the March 2024 MIP amendment removing the 13.5% escalator, which the note says “led to a significant increase in the charge and liability”; the March 2026 recognition of a $774M capital contribution to APIC from the PIK forgiveness (correct under ASC 470-50-40-2, but it took equity from $(964)M to $(237)M without the business earning it); and the FY2025 recognition of a $23.8M UK deferred tax asset “due to the implementation of restructuring,” the largest single component of the $31M tax credit. Separately, the company is mid-rollout of an upgraded ERP system, with $5M of costs added back to adjusted EBITDA — relevant to the disclosed general-IT-controls material weakness. [F]

Recent changes — new markets, facilities, management? New markets: the 15-year partnership takes DPC into aeroengine blades and vanes — a new product category with a large aftermarket tail — from 2029, plus a customer-underwritten greenfield superalloy plant also from 2029. [F] New facilities: approximately $52M at Bochum (a 3,000-square-metre production hall from spring 2026 plus a series of DS furnaces on a 12–18-month installation programme), DS/SX capability at Oxford (Alabama), an end-to-end conversion at Mexicali, and a doubled shell line at Groton — the last of which took production offline during 2025, explaining why aerospace grew only 9.0% in a strong year. [F] Management and board: a refreshed team since the 2020 restructuring — CEO Michael Joseph Quinn (ex-PCC Group VP 2005–2014), CFO David John Egan, COO Jason Mays (Howmet 1990–2006). The board includes Stanley Deal, former President and CEO of Boeing Commercial Airplanes — a genuinely high-quality appointment — alongside Henry F. Brooks, Taiwo K. Danmola, Willibald Meixner, Nicholas Sanders (ex-Gardner Aerospace), Dirkson Charles (co-chair) and C. Alexander Harman (Managing Partner, J.F. Lehman). [F] Corporate: a 4-for-1 reverse share split on 2026-06-05; renamed from Alloy Topco Limited in December 2025 and converted to a Jersey PLC before listing. [F] Capital markets: the IPO launched at ~$700M, was marketed at $28–32, and priced above range at $33.00 on an upsized 27,858,585 shares (2026-06-24); first-day close +42%; four syndicate initiations on 2026-07-20 (Jefferies Buy, Morgan Stanley Equal-weight, RBC Outperform, Rothschild & Co Redburn Buy), with the all-time high of $53.50 exactly matching the highest published target. No Barclays or Moelis note has appeared, leaving two further initiations outstanding. First-ever results are due 2026-08-11 pre-open with a 10:00 ET call. [F]

APPENDIX B — Source Appendix

DPC Holdings Limited (NYSE: DPC) · Report date 2026-07-25 · All sources accessed 2026-07-25 unless stated.

A note on the evidentiary base. DPC has filed no 10-K, no 10-Q and no 8-K, and has never held an earnings call. The 424B4 prospectus is the sole primary source for every financial figure in this article. Because the issuer discloses three material weaknesses in internal control over financial reporting — including an inability to verify that its IFRS-to-US-GAAP conversion adjustments “are complete and accurate” — and because no SOX 404(b) auditor attestation is required for up to two more years, every figure derived from it is provisional. Sources reachable only via a search index (direct fetch returning HTTP 403) are flagged as such and should be re-verified before any material reliance.

Primary — SEC filings (EDGAR, CIK 0002107018)

The complete corpus is 25 filings over 101 days. No filings exist after 2026-06-26.

1. Prospectus (Form 424B4), filed 2026-06-26, prospectus dated 2026-06-24, Registration No. 333-296215 — THE primary source. https://www.sec.gov/Archives/edgar/data/2107018/000110465926077968/tm269965-19_424b4.htm (836,821 characters of extracted text). Sections relied on throughout:
   • Cover page and “The Offering” (pp.1–21) — offering size, pricing, private placements, share counts, reverse split
   • “Summary — Our Company / Competitive Strengths / Growth Strategy” (pp.1–16) — business description, the four strategic partnerships, “up to 80% of customer contribution to capital investment,” customer concentration, LTA terms, backlog
   • “Summary Consolidated Financial and Other Data” (pp.22–24)
   • “Risk Factors” (pp.25–50) — including the three material weaknesses (pp.43–45), single-site concentration, termination-for-convenience, energy costs, labour, tariffs, and the absence of any product-liability/quality-escape risk factor
   • “Use of Proceeds” (p.53) — the itemised $330M / $50M / $155M / $280M allocation
   • “Dividend Policy” (p.54), “Capitalization” (p.55), “Dilution” (p.56)
   • “MD&A” (pp.58–84) — results of operations, segment information (pp.65–69), capital expenditures (p.71), liquidity and capital resources (pp.72–73), contractual obligations and off-balance-sheet arrangements (p.73), “Non-GAAP Financial Measures” (pp.78–81) — the adjusted EBITDA and adjusted net income reconciliations
   • “Business” (pp.85–105) — segments, facilities, customers and suppliers, intellectual property, government and industry regulation, employees, legal proceedings
   • “Management” (p.106), “Executive Compensation” (p.113) — the MIP participant amounts, IPO Grants with escalating strikes, MIP Recognition Grants, and the remuneration-committee composition
   • “Certain Relationships and Related Party Transactions” (p.121) — the Management Incentive Plan, the Concurrent Private Placement, the shareholder transfer agreement
   • “Principal Shareholders” (pp.123–128) — the 5% holder table and footnotes
   • “Description of Share Capital” (p.129), “Shares Eligible for Future Sale” (p.141), “Taxation” (p.143), “Underwriting” (p.151) — the 180-day lock-up and Jefferies/Morgan Stanley waiver discretion, the QIA placement agent fee, the Directed Share Program
   • “Enforcement of Civil Liabilities” (p.163)
   • Audited consolidated financial statements (F-2 to F-7) and Notes — KPMG report (F-2); balance sheet, income statement, comprehensive income, cash flow and equity statements (F-3 to F-7); Note 1 Nature of the business; Note 2 Significant accounting policies (incl. ex-works revenue recognition, F-12); Note 3 Revenue (F-22); Note 4 Segment reporting (F-22 to F-24); Note 5 Income taxes (F-25 to F-27); Notes 7–11 receivables, inventories, PP&E, goodwill and intangibles (F-28 to F-30); Note 12 Leases; Note 13 Borrowings (incl. the Shareholder PIK Loan and the Term Loan springing-maturity condition); Note 15 Accruals and other liabilities (F-35); Note 16 Pensions (F-36); Note 17 Commitments (F-44/45); Note 18 Management Incentive Plan; Note 19 Disposal group held for sale (F-46); Note 21 Share capital
   • Unaudited interim condensed consolidated financial statements for the three months ended 2026-03-29 and 2026-03-30 (F-50 onward), and Notes 4, 8, 9, 10
2. Form S-1/A, filed 2026-06-22 — establishes the marketed range (“between $28.00 and $32.00”), 23,333,333 shares offered, a 3,499,999-share option, ~$782M of estimated net proceeds at the $30.00 midpoint. https://www.sec.gov/Archives/edgar/data/2107018/000110465926076488/tm269965-16_s1a.htm
3. Form S-8, filed 2026-06-25 — registration of equity-plan shares. https://www.sec.gov/Archives/edgar/data/2107018/000110465926077649/tm269965d21_s8.htm
4. Free Writing Prospectus (FWP), filed 2026-06-22. https://www.sec.gov/Archives/edgar/data/2107018/000110465926075947/tm269965d15_fwp.htm
5. Eleven Form 3s, filed 2026-06-24 — initial statements of beneficial ownership, establishing pre-IPO insider holdings of ~0.44% of the company.
6. Nine Form 4s, filed 2026-06-26 — the IPO-related transactions, ~$75.8M of code-P purchases of which ~80% is recycled MIP proceeds. (Note per repo LEARNINGS: SEC Form 4s are mirrored as .xml and must be parsed as XML or read via the human-readable rendering.)
7. Form 8-A12B and NYSE CERT (both 2026-06-24); S-1MEF and EFFECT (2026-06-24); Form S-1 (2026-05-26); DRS (2026-04-15).
8. EDGAR submissions JSON, CIK 0002107018 — https://data.sec.gov/submissions/CIK0002107018.json — confirms registrant name “DPC Holdings Ltd”, SIC 3360 (Nonferrous Foundries — Castings), UK business address (Donington Court, Pegasus Business Park, Herald Way, Derby DE74 2UZ), ticker DPC, and the complete filing history.

Primary — competitor and comparator filings

9. Howmet Aerospace Inc., Form 10-K FY2024, filed 2025-02-14 — https://www.howmet.com/wp-content/uploads/sites/3/2025/10/10-K-2024.pdf — establishes that Howmet is NOT vertically integrated into nickel superalloy melt (“requires nickel, titanium, aluminum, cobalt, and superalloy materials,” purchased on the open market), and names “Doncasters (UK)” among ~seven investment-casting competitors while identifying only PCC as the full-overlap peer.
10. Berkshire Hathaway Inc., 2025 Annual Report and Form 10-K — https://www.berkshirehathaway.com/2025ar/2025ar.pdf and https://www.berkshirehathaway.com/2025ar/202510-k.pdf — PCC segment figures (2024 revenue $10.4B, pre-tax ~$1.9B, 2025 operating cash flow $2.4B). FLAG: the PDFs did not render to text for direct quotation; these figures are taken from search-indexed summaries and should be re-verified before material reliance.
11. Berkshire Hathaway / PCC transaction release, 2015-08-10 — https://www.globenewswire.com/de/news-release/2015/08/10/759226/0/en/Berkshire-Hathaway-Inc-to-Acquire-Precision-Castparts-Corp-for-235-Per-Share-in-Cash.html — the $235/share, ~$32.1B equity / ~$37B EV purchase price.
12. Precision Castparts / Special Metals release, 2005-08-26 — https://www.globenewswire.com/news-release/2005/08/26/332291/2215/en/Precision-Castparts-Corp-to-Acquire-Special-Metals-Corporation-in-Accretive-Transaction.html — establishes that PCC’s nickel-melt vertical integration predates DPC’s by roughly two decades.

Primary — company (Doncasters) and pre-IPO corporate history

13. Dundee Pikco Limited, Annual Report 2019 / s.172 statement — https://www.doncasters.com/wp-content/uploads/2021/03/dundee-pikco-limited-2019-section-172-statement.pdf — the phased divestment programme from December 2017: Fasteners to Stanley Black & Decker (SPA 2017-12-22, completed 2018-04-02, $426.0M cash); two US Fabrications sites (2018-12-21 and 2019-01-10, $8.0M plus a $1.5M earn-out); Settas SA to Precimetal (2019-09-12). “Dundee Pikco Limited” was the prior group holding-company name. FLAG: direct fetch returned HTTP 403; details taken from indexed content.
14. Doncasters Group, “Contract worth over $1 billion USD signed with Siemens Energy,” August 2023 — https://www.doncasters.com/news/contract-worth-over-1-billion-usd-signed-with-siemens-energy — the probable origin of Customer A and of the IGT ramp. FLAG: HTTP 403 on direct fetch; headline, value and scope from the indexed summary.
15. Doncasters Group, “Doncasters announces $52 million investment in Bochum Manufacturing Facility” — https://www.doncasters.com/doncasters-announces-52-million-investment-in-bochum-manufacturing-facility/ — the 3,000 sq m production hall from spring 2026 and the DS furnace series. FLAG: HTTP 403. Cross-referenced against Shop Metalworking Technology — https://shopmetaltech.com/manufacturing/doncasters-announces-52-million-investment-in-bochum-germany-facility/
16. Doncasters Group, “Doncasters Invests In Alabama Facility” — https://www.doncasters.com/news/doncasters-invests-in-alabama-facility — ~$3.5M for a 300lb equiaxed vacuum casting furnace at Oxford, Alabama. FLAG: indexed summary.
17. Latham & Watkins, “Latham Advises Ad Hoc Group of First-Lien Lenders on Doncasters Group Financial Restructuring,” March 2020 — https://www.lw.com/en/news/2020/03/latham-advises-lenders-on-doncasters-restructuring
18. Weil, Gotshal & Manges, “Doncasters’ Restructuring Successfully Implemented” — https://www.weil.com/\~/media/weil-london-thought-leadership/doncasters_successfully_implements_restructuring.pdf
19. TheBusinessDesk.com, “Restructuring sees £900m of debt wiped from 240-year-old manufacturer” — https://www.thebusinessdesk.com/eastmidlands/news/2035734-restructuring-sees-900m-of-debt-wiped-from-240-year-old-manufacturer — £1.22B of funded debt restructured 2020-03-06; ~£900M written off; new CEO and non-executive chairman; ownership transferred to the funders; prior owner Dubai International Capital.
20. Bloomberg Law, “Doncasters Group’s Creditors Seeking Talks on Restructuring” (2019) — https://news.bloomberglaw.com/bankruptcy-law/doncasters-groups-creditors-seeking-talks-on-restructuring — the $745M secured facility maturing April 2020; PJT Partners and Latham & Watkins retained by lenders.

Primary — IPO transaction and post-IPO events

21. Businesswire, “Doncasters Announces Pricing of Upsized Initial Public Offering,” 2026-06-24 — https://www.businesswire.com/news/home/20260624379322/en/Doncasters-Announces-Pricing-of-Upsized-Initial-Public-Offering
22. Businesswire, “DPC Holdings Announces Date and Time for Second Quarter 2026 Financial Results and Webcast,” 2026-07-20 — https://www.businesswire.com/news/home/20260720988038/en/DPC-Holdings-Announces-Date-and-Time-for-Second-Quarter-2026-Financial-Results-and-Webcast — Q2 2026 results 2026-08-11 pre-open, 10:00 ET call.
23. Davis Polk & Wardwell (deal counsel), “DPC Holdings $1 billion IPO” — https://www.davispolk.com/experience/dpc-holdings-1-billion-ipo — describes “32,037,372 ordinary shares … which included 4,178,787 ordinary shares sold pursuant to the underwriters’ over-allotment option,” the basis for concluding the greenshoe was exercised in full.
24. White & Case LLP, “White & Case advises Doncasters on upsized US$1.1 billion IPO” — https://www.whitecase.com/news/press-release/white-case-advises-doncasters-upsized-us11-billion-ipo — FLAG: HTTP 403 on automated fetch; cited via search summary.
25. Reuters / Investing.com, “Aerospace parts maker Doncasters raises $919.3 million in US IPO,” 2026-06-25 — https://m.investing.com/news/stock-market-news/aerospace-parts-maker-doncasters-raises-9193million-in-us-ipo-4759700
26. IPOScoop, “The IPO Buzz: Doncasters – DPC Holdings Prices Upsized IPO at $33 – Above Range & Stock Jumps,” 2026-06-25 — https://www.iposcoop.com/the-ipo-buzz-doncasters-dpc-holdings-dpc-upsizes-ipo-prices-it-at-33-1-above-range/
27. Renaissance Capital IPO Center, “Aerospace and defense engine manufacturer Doncasters Group sets terms for $700 million US IPO” — https://www.renaissancecapital.com/IPO-Center/News/119824/ — establishes the deal’s original ~$700M size before upsizing.
28. Sell-side initiations, all dated 2026-07-20 — reported as positioning evidence only; no third-party price target is adopted as our view. Morgan Stanley Equal-weight (analyst Kristine Liwag) — https://m.investing.com/news/analyst-ratings/morgan-stanley-initiates-dpc-holdings-stock-at-equalweight-47-target-93CH-4799868 ; Jefferies Buy — https://www.investing.com/news/analyst-ratings/jefferies-initiates-dpc-holdings-stock-coverage-with-buy-rating-93CH-4800267 ; Rothschild & Co Redburn Buy — https://www.investing.com/news/analyst-ratings/rothschild-redburn-initiates-dpc-holdings-stock-with-buy-rating-93CH-4800016 ; RBC Capital Markets Outperform — via https://stockanalysis.com/stocks/dpc/
29. NYSE / PR Newswire pre-market advisory, “Doncasters Shares Jump 42% in NYSE Trading Debut,” 2026-06-26.
30. Seeking Alpha: “DPC Holdings: Flying Higher Out Of The Gate” (2026-06-26) and “DPC Holdings: Visible Customer-Funded Growth” (2026-07-23) — the only two independent published analyses in existence.

Industry — capacity, demand and the capital cycle

31. GE Aerospace Q2 2026 results (revenue +24%, EPS $2.02 vs $1.86 expected, FY26 revenue/EPS/FCF guidance raised, Commercial Engine Services orders +18%) — https://www.investing.com/news/transcripts/earnings-call-transcript-ge-aerospace-beats-q2-2026-estimates-raises-outlook-93CH-4795780 and https://leehamnews.com/2026/07/16/q2-2026-earnings-ge-aerospace-keeps-the-beat-despite-a-few-missed-notes/ — the identifiable sector catalyst for DPC’s 2026-07-21→24 move.
32. Howmet capacity expansion commentary (capex to ~$470–500M FY2026, ~70% into Engine Products, IGT capital deals closed with six of seven customers) — https://finance.yahoo.com/markets/stocks/articles/howmet-hwm-speaks-capacity-expansion-191650404.html
33. Precision Castparts / Morvern Group acquisition, closed 2026-03-01 — https://www.mainsights.io/ma-news/us-based-metal-fabrication-firm-precision-castparts-acquires-uk-aerospace-and-defence-engineering-company-morvern-group and https://aviationoutlook.substack.com/p/precision-castparts-company-analysis-outlook-report — PCC’s first acquisition since the Berkshire takeover; wax patterns, ceramic core assemblies and tooling; Derby and Worcester; ~£22M FY2025 revenue.
34. Consolidated Precision Products / Warburg Pincus and Berkshire Partners — https://warburgpincus.com/investments/consolidated-precision-products/ and https://berkshirepartners.com/consolidated-precision-products-corp-announces-recapitalization-with-berkshire-partners-and-warburg-pincus/ — CPP ownership and the Euclid, Ohio expansion.
35. Chromalloy FAA PMA on CFM high-pressure-turbine blades, with a US commercial airline launch customer — https://www.asminternational.org/tss/chromalloy-secures-faa-approval-of-cfm56-high-pressure-turbine-blade-pma/ and https://www.accessnewswire.com/newsroom/en/aerospace-and-defense/u.s.-commercial-airline-selects-chromalloy-cfm-high-pressure-turbine-blade-pm-1159189
36. PCC post-write-down downsizing — https://www.newequipment.com/plant-operations/article/21138901/after-writedown-precision-castparts-is-downsizing-berkshire-hathaway — the forging, casting and fastener business “reduced total employment by 30%, or about 10,000, during the first half of 2020” off a 33,417 year-end-2019 base; Berkshire’s $9.8B Q2-2020 write-down.
37. Ceramic matrix composites — the substitution threat. GE Aerospace: https://www.geaerospace.com/news/press-releases/defense-engines/ge-successfully-tests-worlds-first-rotating-ceramic-matrix-composite and https://www.geaerospace.com/news/articles/100-year-anniversary-paris-airshow-technology/ceramic-matrix-composite-technology-ges ; ACerS Bulletin: https://bulletin.ceramics.org/article/ceramic-matrix-composites-taking-flight-at-ge-aviation/ ; ORNL: https://www.ornl.gov/news/ceramic-matrix-composites-take-flight-leap-jet-engine — GE9X carries >100 CMC parts across five hot-section components; HPT shrouds at Manufacturing Readiness Level 10 (full-rate production).
38. Additive manufacturing is not yet a substitute for DS/SX airfoils — https://www.sciencedirect.com/science/article/pii/S2950431726000043 ; https://pure.psu.edu/en/publications/on-the-fabrication-of-metallic-single-crystal-turbine-blades-with/ ; https://www.additivemanufacturing.media/articles/3d-printed-turbine-blades-for-more-efficient-power-generation-the-cool-parts-show-35 ; https://www.prodways.com/resources/single-crystal-casting-3d-printing/
39. Input costs. Cobalt (~$56,414/t entering 2026, highest since July 2022; DRC export quotas capping hydroxide at ~96,000t for 2026) — https://investingnews.com/daily/resource-investing/battery-metals-investing/cobalt-investing/cobalt-forecast/ and https://www.canadianminingreport.com/blog/cobalt-market-outlook-2026-key-drivers-and-risks ; Indonesia nickel (60.2% of global output; 2026 quota cut to 260–270Mt from 379Mt) — https://www.ebc.com/forex/one-country-60-of-global-nickel-and-a-30-cut-the-shockwave-hitting-defense-evs-and-global-economy ; LME nickel — https://www.tacto.ai/en/commodities/nickel-price
40. Tariffs. Federal Register / EO 14309 implementation — https://www.federalregister.gov/documents/2025/06/30/2025-12060/imports-of-automobiles-automobile-parts-civil-aircraft-and-civil-aircraft-parts-from-the-united ; White & Case on Section 232 modifications — https://www.whitecase.com/insight-alert/united-states-modifies-steel-aluminum-and-copper-section-232-tariffs ; the 2026-07-09 Section 232 aerospace proclamation imposing no new tariffs but mandating negotiations within 180 days — https://www.internationaltradeinsights.com/2026/07/trump-administration-declines-to-institute-section-232-tariffs-on-commercial-aircrafts-jet-engines-and-aircraft-parts/ and https://www.csis.org/analysis/section-232-investigation-commercial-aircraft ; Make UK — https://www.makeuk.org/tariffs
41. European industrial energy costs. House of Lords Library — https://lordslibrary.parliament.uk/electricity-prices-in-great-britain/ ; UK Steel — https://www.uksteel.org/electricity-prices ; IEA Electricity 2026 — https://www.iea.org/reports/electricity-2026/prices ; Nesta — https://www.nesta.org.uk/blog/how-could-the-uk-lower-its-industrial-electricity-prices/

Third-party quantitative data (not primary — reconciled to filings where material)

42. AZI Trading price history CSV, DPC and comparators (HWM, CRS, ATI, HEI, TDG, MOG-A, KRMN) — https://azitrading.com/controls/download-data.php?t=DPC — 21 sessions from 2026-06-25; Source of the $49.91 close, the $53.50 high, the $42.90 low, volumes and realized volatility. Note: the CSV’s beta and alpha columns are small-sample artefacts on a 21-observation series and are not reported as facts; the 21/50/200 EMAs are all seeded from the day-one close and carry no information.
43. AZI Trading ticker page — https://azitrading.com/DPC — sector Industrials, industry Aerospace & Defense.
44. AZI fundamentals feed — valuation_index.latest = null, history = null. The stock’s own-history valuation percentile, normally this firm’s highest-signal valuation datum, does not exist for DPC. Accordingly no cross-sectional substitute was used.
45. FactorsToday — https://www.factorstoday.com/api — returns no factor data for DPC: /stock-loadings/DPC = [], /leaderboard/DPC = {"data":null}, /stock-specific-vol/DPC = error, /related-stocks/DPC = []. /stock-info/DPC confirms the cause: trading_history = 21 against a 252-day universe minimum. No loadings were fabricated. Comparator loadings, leaderboards and /factor-returns/historic were pulled for HWM, CRS, ATI, MOG-A, HEI, TDG, GE and GEV as the substitute positioning read. Methodology: https://www.factorstoday.com/about
46. ROIC.ai MCP — get_company_profile (ISIN JE00BPK4NV88, Jersey; CIK 0002107018; NYSE; CEO Michael Joseph Quinn; 3,070 full-time employees; IPO date 2026-06-25; is_adr = false), get_company_news (9 articles since 2026-04-01 — the complete post-IPO tape, all IPO mechanics and initiations, zero operational news), and statement/ratio tools. Two errors flagged: ROIC.ai reports FY2025 EPS of $(1.19) and FY2024 of $(1.33) by retroactively applying 145,151,799 shares; the filing’s $(1.53) and $(1.71) on 112,936,894 weighted-average shares are authoritative. Third-party aggregated data, not primary.
47. yfinance — cross-sectional comparator multiples for HWM, CRS, ATI, HEI, TDG, GE, RTX, MOG-A, KRMN, CW, WWD, GEV, SIEGY, RYCEY and SAF.PA. UNOFFICIAL. Note per repo LEARNINGS: yfinance’s DPC enterprise value of $6.34B is wrong because it applies a pre-IPO balance sheet; all DPC balance-sheet figures in this article come from the prospectus.
48. stockanalysis.com/stocks/dpc — share count, aggregated analyst ratings and targets. Note: third-party sites quoting a “52-week range of $30.00–$53.50” are wrong on the low — the traded low is $42.90 and $30.00 was never a traded price.

Notes on data reliability

• The issuer’s own internal controls are disclosed as deficient in three respects, one of which is the inability to verify the completeness and accuracy of its IFRS-to-US-GAAP conversion adjustments. No SOX 404(b) attestation is required for up to two more years. Every financial figure in this article should be treated as provisional pending the first 10-Q (2026-08-11).
• The share count is unresolved. The prospectus states 145,151,799 shares outstanding post-offering; deal counsel’s description of a 32,037,372-share offering including the full 4,178,787-share over-allotment implies 149,330,586. No 8-K was filed. Every third-party feed still shows 145.15M or fewer — FactorsToday’s implied 142.24M is wrong by ~2%. The 10-Q cover on 2026-08-11 is authoritative. The difference moves market capitalisation and every multiple by roughly 1–3%.
• Three prospectus figures are internally inconsistent and were resolved in favour of the audited notes: “DPC New England” is missing from the 14-site facilities table; Ivostud is described under Engine Products — Europe in one place and under Turbo Wheels in Note 4; engineer headcount appears as both “385” and “380”; the accreditation standard appears as both “AS 9001” and “AS9100”; top-two customer concentration is described as “>30%” against actual figures of 38% (FY2025) and 34% (FY2024); and HIP, X-ray and heat treatment are described as in-house in the Business section and as outsourced in the Risk Factors.
• Six items are disclosed nowhere and are therefore estimated or left open, and each is flagged where used: the amount and accounting of customer capital contributions; stock-based compensation expense and grant-date fair value; FY2019 revenue on the current perimeter; constant-currency revenue growth; debt covenant levels and headroom; and environmental or occupational-disease reserves (the words “asbestos,” “silica,” “beryllium” and “hexavalent” each appear zero times across the 836KB document).
• The prospectus contains zero mentions of “composite” or “CMC”, despite ceramic matrix composites being in full-rate production on GE9X high-pressure-turbine shrouds — the same static hot-section content DPC casts. Verified by full-text search.

This is an independent research article written for general information. It is not investment advice, not a solicitation, and not a recommendation to buy or sell any security. The numbered sections carry no recommendation and no price target; the only exception is the clearly-labelled opinion block at the top, which is the author’s own subjective view. The author may or may not hold a position in any security mentioned. Do your own research.