Carpenter Technology Corporation (NYSE: CRS) — A Superb Cyclical Wearing a Secular Price Tag

Independent equity research — for general information only. Date: 2026-06-19 · Price (2026-06-18 close): $586.47 · Shares: ~49.7M · Market cap: ~$29.0B · Net debt: ~$0.38B · EV: ~$29.4B Fiscal year: ends June 30 · CIK: 0000017843 · Sector: Materials — Specialty Steel & High-Performance Alloys

⚡ Claude’s Take

This block is the author’s own independent opinion. It is general information, not investment advice. The analysis that follows (sections 1–15) is presented position-free and carries no price target, by design.

Verdict: HOLD / AVOID-at-this-price / NOT-a-short. A genuinely excellent, structurally improved business at the richest valuation in its own history — a great company at a demanding price, not a great investment at $586. Conviction: medium-high on the business quality; medium on the call (the multiple, not the franchise, is the entire debate).

Carpenter is the real thing: one of roughly three qualified Western producers of premium vacuum-melted nickel-cobalt superalloys, a node gated by AMS qualification, multi-year requalification switching costs, and capital-intensive VIM/VAR/ESR melt capacity that nobody is building greenfield. The supply side has consolidated (Haynes→Acerinox Nov-2024, Universal Stainless→Aperam Jan-2025), demand is stacking (narrowbody engine ramp + an early-innings spares supercycle + defense + data-center-driven industrial gas turbines), and management has executed beautifully — ROIC from negative in FY21–22 to 17.2% in FY25, ex-surcharge SAO operating margin to a record 35.6% in the latest quarter, a clean balance sheet (0.5x net leverage), and a high-return Athens expansion funded entirely internally. This is a textbook Marathon “favorable supply side” set-up where elevated returns are not attracting the capacity that would compete them away.

But the market knows all of it. CRS trades at the 96.8th percentile of its own ten-year valuation range (P/B and P/S literally at the 99.98th percentile — richest ever), ~62x trailing earnings, ~42x EV/EBITDA, 12.5x book — above ATI and roughly in line with the higher-quality, more-recurring Howmet. That is a peak multiple on peak-ish, still-cyclical margins — the classic double-count. The distribution is negatively skewed: even if operations stay superb (the base case), multiple normalization caps the return; a cycle wobble is a 40–60% drawdown against a business with a −85% lifetime maximum drawdown and ~0% gross margin as recently as FY21. The framing is crowded high-beta aerospace momentum (market beta 1.5, A&D factor loading 0.78, +134% one-year return), not falling knife and not value. Tag: right alloy, wrong entry. I would own the franchise enthusiastically at a price that respects its cyclicality — a re-rating toward ~18–22x mid-cycle EBITDA (roughly a high-$200s to high-$300s zone, depending on where one pegs normalized EBITDA) is where risk/reward turns genuinely attractive; at $586 I am a patient watcher, not a buyer, and not a short into a working momentum trade.

• Flips bullish: the FY27 guidance reset (Q4 FY26 call, ~late July 2026) plus new long-term targets confirm ex-surcharge margins durably stepping past ~24% and the multiple compressing on a pullback into the zone above — i.e., the structural-baseline case proven while the price de-risks.
• Flips bearish (short-worthy): a Boeing/Airbus build-rate cut, an A&D supply-chain destock, or a momentum unwind while the stock still carries a 40x+ EV/EBITDA multiple — earnings and multiple rolling together.

📈 Stock Price Action — Five-Year Event Map

Factual price history, not a recommendation. Price moves are FACT; attributed drivers are INTERPRETATION. No price target, no support/resistance levels.

The arc. CRS round-tripped from a COVID trough of $13.69 (2020-03-23) through a post-COVID false-start washout low of $24.59 (2022-07-11), based for most of 2022, then climbed relentlessly: ~$43→$71 across 2023, a vertical melt-up to ~$194 by late 2024, a consolidation into mid-2025 (~$175–240), and a renewed acceleration to an all-time high of $586 (2026-06-18). The 52-week range is roughly $230 (2025-09-25) to $586 — the stock sits at its all-time high, ~58% above its 200-day EMA (~$371). Over the trailing 60 months CRS is up ~22x off the 2022 low and ~2.5x in the last twelve months alone.

Narrative. CRS is a ~22-bagger off the 2022 low powered by both a real fundamental turn (ROIC negative FY21–22 → 17.2% FY25; EPS $7.42 FY25 → ~$10.43 FY26E) and a violent multiple re-rating from distressed-cyclical (~10–15x earnings in 2022) to ~50–62x today. Legs 3–4 were fundamentals catching up to a depressed price; legs 5, 7 and 8 are increasingly multiple-and-momentum driven on top of genuinely rising earnings. Each up-leg ties to a specific earnings print or guidance raise cross-referenced to the FY26 nine-month result (+40% EPS) and consensus revisions; the price facts come from the AZI five-year daily series.

1. Executive Summary

Carpenter Technology is a 135-year-old Philadelphia-based producer of premium specialty alloys — high-performance nickel and cobalt superalloys, titanium, stainless and tool steels, plus metal powders and additive parts — sold primarily into aerospace & defense (~62% of revenue and rising), with medical, energy/industrial-gas-turbine, transportation and industrial end markets. It operates two segments: Specialty Alloys Operations (SAO), the melt-and-mill franchise that generates ~94% of segment operating income, and the much smaller Performance Engineered Products (PEP) (Dynamet titanium, Carpenter Additive powder/AM, Latrobe distribution).

The investment story is a genuine, well-executed cyclical recovery that has matured into a momentum re-rating. After a near-death COVID experience — FY21 gross margin of 0.07%, an operating loss, ROE of −13.4%, EPS of −$4.75 — Carpenter has staged one of the cleaner industrial turnarounds in the market: revenue $1.48B (FY21) → $2.88B (FY25); gross margin to 26.7% (32.8% ex-surcharge); operating margin to 18.3%; ROIC from negative to 17.2%; diluted EPS to $7.42 (FY25) and a trailing-twelve-month $9.51. FY26 is tracking ~+40% YoY (nine-month diluted EPS $7.30 vs $5.21), with the most recent quarter a record (SAO operating margin 35.6%, the 17th consecutive quarter of margin expansion). Management has raised FY26 operating-income guidance three times to $700–705M (“at least +33%”) and has signaled it will reset its FY27 target upward and add longer-term guidance on the Q4 FY26 call — a near-term catalyst.

The business quality is real. The moat is intangibles (AMS/customer qualification incumbency) plus switching costs, reinforced by niche scale in capital-intensive vacuum melt — durable but narrow, and one value-chain node below Howmet’s cast-part franchise. The industry structure is favorable and consolidating, with no net new qualified Western capacity since 2019 and Carpenter’s own ~7% brownfield Athens add deliberately too small to break pricing. Capital allocation has been disciplined (balance sheet repaired, dividend protected not hiked, modest buyback restarted only after net leverage fell below 1x, ~37% pre-tax-return Athens expansion self-funded, no value-destructive M&A).

The problem is price. CRS trades at the 96.8th percentile of its own decade-long valuation range — P/B and P/S at the 99.98th percentile (richest ever), ~62x trailing earnings, ~42x EV/EBITDA, 12.5x book — above ATI and roughly in line with the higher-quality Howmet. That is a peak multiple applied to peak-ish, still-cyclical margins. The embedded expectation is “secular compounder,” not “cyclical at peak.” The distribution of forward outcomes is negatively skewed: even the base case (operations stay excellent) yields a poor return as the multiple normalizes, while a cycle roll re-rates earnings and multiple together against a business with a −85% lifetime drawdown history. This memo takes no position and sets no price target (the one labeled exception is Claude’s Take above). It argues that the central, and largely only, question for forward returns from $586 is whether a record cyclical multiple on record cyclical margins can persist.

2. Business Overview

What Carpenter does. Carpenter is a premium specialty-alloy melter and mill-products producer. It melts and processes high-performance metals — nickel- and cobalt-based superalloys, titanium alloys, stainless steels, alloy and tool steels, soft magnetic alloys, powder metals and additive parts — into billet, bar, rod, wire and strip, plus powders and near-net-shape parts. Crucially, Carpenter sits one node upstream of the visible aerospace component makers: its mill products are the qualified input that forgers and casters such as Howmet, Precision Castparts and the engine OEMs machine into rotating engine parts, fasteners, structural components, implants and tooling. It is, in essence, the metallurgy layer of the aerospace and medical supply chains.

Two segments — extremely lopsided economics.

• Specialty Alloys Operations (SAO) is the core franchise: integrated premium melt (VIM primary melt plus VAR/ESR remelt), hot/cold processing, and finishing across the Reading PA, Latrobe PA and (under construction) Athens AL footprint, with a global service-center distribution network. In FY25, SAO produced roughly $588.6M of operating income at a ~23.0% reported margin (28.6% ex-surcharge) — about 94% of pre-corporate segment operating income.
• Performance Engineered Products (PEP) bundles Dynamet (titanium bar/wire, heavily medical-distribution-exposed), Carpenter Additive (metal powder + additive manufacturing) and Latrobe distribution. FY25 PEP operating income was ~$37.0M at a ~9.1% margin — about 6% of segment profit. Carpenter Additive was a value-destructive 2017–18 acquisition spree (Puris, LPW) that required restructuring and impairment, and is only now turning into a small “bright spot.”

The segment bridge: SAO $588.6M + PEP $37.0M − corporate $102.9M ≈ $521.8M consolidated operating income (FY25). The investment case is therefore, to a first approximation, a bet on SAO — premium aerospace melt.

End markets. Aerospace & defense is ~62% of revenue and rising (management expects >60% to keep climbing as A&D outgrows the rest); medical ~12%; energy ~7% (increasingly industrial-gas-turbine / data-center-power driven); the remainder transportation, industrial and consumer. Within A&D, the engine sub-market (rotating-part superalloys, fasteners) is the highest-value, stickiest demand; aero-structural is lower-MRO-content and the swing factor that destocked hardest on Boeing’s 737 issues and is now “moving off the sidelines.”

How it makes money, and the surcharge distortion. Reported revenue includes a raw-material surcharge pass-through — roughly $531M of FY25’s $2,877M revenue is near-zero-margin surcharge that mechanically passes nickel/cobalt/titanium cost changes to customers. This is the single most important thing to understand about Carpenter’s financial optics: all growth and margin analysis must be done on the ex-surcharge basis (FY25 ex-surcharge sales $2,346M). The surcharge dilutes the percentage margin (it adds revenue at ~0% margin), so reported gross margin of 26.7% understates the true 32.8% ex-surcharge gross margin. Mechanically it protects gross-profit dollars against metal-price swings but introduces a quarter-level timing lag (the formula uses prior-period published metal prices), which is a symmetric modeling caveat, not a one-way flatter.

Recurring vs. project. Roughly 40% of total revenue is under long-term agreements (LTAs), rising to ~60–65% of aerospace revenue; the balance is transactional — but transactional customers are long-standing and highly specified, and management notes transactional pricing typically exceeds LTA pricing. Demand is “recurring” in the sense of multi-year qualified positions on long-life engine/airframe/implant programs, but it is cyclical recurring — tied to OEM build rates and aftermarket shop-visit cadence, not subscription-like.

Verdict (Business Overview): A focused, premium-tier specialty-alloy melter with genuinely attractive SAO economics and a small, historically troubled PEP. The business is far higher-quality than the commodity-steel label implies, but its profit is concentrated in one cyclical, aerospace-driven segment, and its reported numbers require the ex-surcharge lens to read honestly.

3. Industry Dynamics

Structure: a tight, consolidating premium-melt oligopoly. The market that matters — qualified, vacuum-melted nickel/cobalt superalloys (and high-end titanium and stainless) for flight-critical aerospace, medical implants and high-temperature energy applications — is a fewer-than-ten-player oligopoly. Carpenter’s own 10-K describes “less than ten companies” as major competitors for its high-value products, against “several dozen” smaller producers and “several hundred” distributors in the commodity tier it deliberately exits. The premium tier passes the Greenwald “count the leaders on one hand” test.

The scarce node is premium melt. Flight-critical rotating-part grades are always vacuum melted — VIM (vacuum induction melting) for the primary melt, then VAR (vacuum arc remelt) and/or ESR (electroslag remelt) — the VIM-VAR route mandated by AMS (Aerospace Material Specifications). This vacuum/remelt capacity is the genuine bottleneck: capital-intensive, long-lead, and qualified furnace-by-furnace, grade-by-grade. Carpenter operates seven qualified VIM furnaces — a meaningful share of Western premium capacity.

Supply is structurally constrained — the heart of the bull case. Three facts compound:

1. No greenfield mills. No new greenfield premium-melt mills have been built in the West in decades. Capacity adds are brownfield — Carpenter’s own ~$400M Athens project adds one VIM furnace, ~9,000 tons (~7% over FY19 shipments), commissioning early FY28, with most customer qualifications not complete until ~CY2030. A new furnace takes years to build and years to requalify per customer/grade.
2. Qualification, not steel-making, is the binding constraint. New capacity does not become sellable premium product for years.
3. Consolidation, not fragmentation. In under twelve months, two independent US pure-plays disappeared: Haynes International (Kokomo nickel/cobalt superalloys) → Acerinox/North American Stainless (closed Nov-21-2024, ~$970M EV), and Universal Stainless & Alloy Products → Aperam (closed Jan-23-2025, ~$537M EV). The field of public pure-plays shrank to Carpenter + ATI.

Demand stack — multi-year and partly uncorrelated. (i) Narrowbody engine ramp: CFM LEAP + Pratt GTF; industry estimates the LEAP+GTF ramp alone adds ~1,300–1,700 tonnes/yr of incremental superalloy demand by 2028. (ii) Engine spares/aftermarket: an aging in-service fleet plus hotter-running new-tech engines (more frequent hot-section replacement) — an early-innings shop-visit supercycle. (iii) GTF/LEAP durability issues are a double-edged driver — they constrain OEM build rates but increase replacement-part demand, a net positive for material suppliers. (iv) Industrial gas turbines / data-center power: the same superalloy technology as jet engines, with GE Vernova / Siemens Energy / Mitsubishi sold out on AI-grid demand — a second, partly-uncorrelated secular leg at aero-like margins. (v) Medical implantables (biocompatible Ti/Co-Cr) — high-value, less cyclical, currently in a destocking trough. (vi) Defense — program-specific, steady-to-rising with global rearmament. (vii) Western titanium tightness post-VSMPO-AVISMA (Russia) sanctions — a tailwind for Carpenter’s Dynamet Ti.

Marathon capital-cycle read — strongly favorable supply side. High and rising returns are not attracting the capacity that would compete them away, because qualification and capital barriers throttle the supply response. There is no flood of new entrants, no IPO wave, and M&A is consolidation by larger strategics, not bubble-priced empire-building. This is exactly the configuration under which elevated returns can persist longer than a normal industrial cycle. The risk lives on the demand end (build-rate stumbles, destocks, recession), not the supply end.

Verdict (Industry): Structurally GOOD — for the premium-melt tier. A consolidating, qualification-gated oligopoly with disciplined, brownfield-only supply against a multi-year secular demand stack. Two qualifiers: this verdict applies to the premium tier only (commodity specialty steel is a bad, fragmented business Carpenter is exiting), and the whole sector remains demand-cyclical — “good industry” here means good through-cycle structure, not low cyclicality.

4. Competitive Position

Name the moat (Greenwald). Carpenter’s moat is intangibles — AMS/customer qualification incumbency — plus switching costs, reinforced by niche economies of scale in premium VIM/VAR/ESR melt. The mechanism:

1. Qualification lock-in / switching costs. The 10-K states products are “generally required to meet complex customer product specifications and often require the materials to be qualified prior to supplying,” and that “experience, technical capabilities, product offerings and R&D … represent barriers to existing and potential competitors.” A specific grade, melted on a specific furnace route, is qualified onto an engine/airframe/implant program. Requalifying a second source costs years plus flight-safety/regulatory risk the OEM will not take absent a quality or capacity failure. This produces sole/limited-source positions and quasi-locked, multi-year demand on long-life programs — the same mechanism as Howmet, one node upstream.
2. Scale in premium melt. Seven qualified VIM furnaces plus remelt, integrated finishing and a global service-center network form a fixed-cost-heavy, hard-to-replicate asset base. A marginal entrant must build capital-intensive vacuum capacity and climb a multi-year qualification curve before earning a dollar. Greenwald’s scale-as-barrier works here because it is combined with customer captivity — entrants cannot simply buy the customers.
3. Metallurgical know-how / process IP. 135 years of melt practice, hundreds of proprietary grades. But — see the pressure-test — R&D is only ~$26M (~1.1% of ex-surcharge sales), and the 10-K notes the business is “not materially dependent upon any single” patent. The durable barrier is qualification/yield/trust, not patents.

Does the moat show up in the financials? (The Greenwald test.) Yes, at the top of the cycle — but the through-cycle record is the catch. ROIC ran 5.0% (FY23) → 14.1% (FY24) → 17.2% (FY25), and SAO ex-surcharge operating margin climbed to a record ~35.6% in a recent FY26 quarter. These are moat-consistent returns near a cycle peak. The problem: FY21 printed a 0.07% gross margin, an operating loss, ROE −13.4%, EPS −$4.75, and a −85% lifetime maximum drawdown. A genuine wide-moat business does not print ~0% gross margin in a demand trough. Howmet stayed solidly profitable through COVID; Carpenter did not. The moat is real but thin and cyclical relative to Howmet’s, because Carpenter sits on the more-commoditizable mill-product node, where pricing power evaporates when aero volume collapses and fixed costs go unabsorbed.

Direct comparison vs. the peer set.

• ATI Inc. (the closest pure comp; 0.96 factor similarity) — larger, more diversified, more vertically integrated downstream into components/forgings. The reference rival; A&D ~68% of revenue (>70% expected).
• Haynes (now Acerinox-owned) and Universal Stainless (now Aperam-owned) — formerly independent, now captive to deep-pocketed strategic parents.
• Precision Castparts (Berkshire) — a vertically integrated caster/forger that also melts (Special Metals, TIMET) — customer and competitor; private, run for cash.
• Howmet (HWM) — a different node: investment castings/forgings/fasteners, a peer-in-value-chain more than a head-to-head melt competitor, and the higher-quality franchise (it stayed profitable through COVID; >50% gas-turbine-blade share).
• VDM/Aperam-Imphy and Aubert & Duval (Europe) — qualified nickel-alloy and Ti/superalloy supply.

Moat-quality ranking: HWM > ATI ≈ CRS > captive Haynes/Universal. Carpenter is the best-positioned of a small oligopoly, not a monopolist, on a value-chain node structurally more cyclical than the cast-component node above it.

Verdict (Competitive Position): A durable advantage — narrow and cyclical, not a wide moat and not a crowded commodity. Intangibles (qualification incumbency) + switching costs + niche melt scale are genuine, show up in sticky qualified positions and rising ROIC, and would clearly cause economics to deteriorate if removed. But the level of returns is heavily operating-leverage-driven and the franchise is one node below Howmet’s, with a far more cyclical history. Durable advantage, medium width, high cyclicality.

5. Growth History and Forward Opportunities

History — a violent V, mostly cyclical. Revenue: $2.18B (FY20) → $1.48B (FY21) COVID trough → $1.84B → $2.55B → $2.76B → $2.88B (FY25). On the cleaner ex-surcharge basis, FY25 sales rose ~8% — but on ~6% lower volume, i.e., the growth was pure price/mix, not units. That distinction matters: Carpenter’s recent revenue growth is dominated by repricing post-COVID LTAs (new agreements at “substantial,” >30% increases; aero price/lb up ~10% YoY) and by mix-shifting toward higher-value A&D grades, not by shipping more pounds. This is high-quality price growth, but it is also a reminder that the volume leg of the story is still ahead — and depends on the aero-structural “bullwhip” actually landing.

Quality of growth. Organic, no acquisitions since 2018. The growth is concentrated in SAO and within SAO in engine and energy/IGT. FY26 quarterly detail: Q3 engine sales +44% YoY, fasteners +20% YoY, energy +44% YoY (“almost 100%” IGT-driven). A&D bookings have accelerated for four-plus consecutive quarters; September-2025 was the highest order-intake month in over a year. The soft spot is medical (Q3 −29% YoY), isolated to certain titanium distribution customers within PEP/Dynamet (a post-COVID destock running longer than expected) — immaterial to total Carpenter and showing booking recovery.

Forward opportunities.

1. Aero-structural restocking. The submarket most hurt by Boeing 737 build-rate issues is “off the sidelines,” placing first large orders in several quarters, with management signaling next orders “larger and more urgent.” This is the volume upside lever — Carpenter is not sold out on structural flow paths (unlike engine).
2. Build-rate ramp. Boeing building 42/mo (Dec-2025), FAA-approved to 42 (Oct-2025), targeting 47 then 52+; Airbus A320neo ramping with further runway. Industry targeting 2,100+ planes/yr, ~30% above the 2019 level at which the nickel-superalloy market was effectively sold out. Engine OEMs report the supply chain “has not ordered enough material” to support targeted rates — the bullwhip.
3. Energy/IGT and data-center power — a strategic, aero-margin growth lever at the same scarce assets.
4. Athens capacity — ~9,000 incremental tons of pre-sold, high-margin premium melt from ~FY28.
5. Defense — multi-platform, elevated, with FY26 budget and geopolitical upside not yet fully felt.

Verdict (Growth): High-quality but cyclically-charged. The recovery growth was genuine and well-executed (price/mix + margin), and the forward demand stack is real and multi-year. But it is concentrated in cyclical A&D, the recent leg was price not volume, and the bull case increasingly leans on a volume ramp (aero-structural restock + build-rate increases) that is forecast, not yet banked. High-quality growth; not low-risk growth.

6. Financial Quality

Margins and operating leverage. The margin recovery is extraordinary: gross margin 0.07% (FY21) → 8.2% → 13.2% → 21.2% → 26.7% (FY25); ex-surcharge gross margin 32.8%; operating margin to 18.3% reported / 22.4% ex-surcharge adjusted; EBITDA margin 23.1%. Incremental operating margins ran ~105% (FY24) and ~146% (FY25) as recovering volume and price/mix flowed over a fixed melt-and-mill cost base near ~80% utilization. SAO ex-surcharge margin reached a record 35.6% in a recent quarter — the 17th consecutive quarter of expansion, with ~80%+ ex-surcharge incremental margins.

Quality of earnings — high-quality, arguably understated. Three points:

• The surcharge works against flattering — it adds ~$531M of near-zero-margin revenue that dilutes the percentage margin. The ex-surcharge adjusted margins (32.8% gross / 22.4% operating) are the true read, and they are strong and rising on genuine mix and pricing.
• The “non-operating income” premise is inverted. Carpenter reports Other expense, net — a net expense ($6.1M FY25, $60.5M FY24). The FY24 figure was inflated by a $51.9M noncash pension settlement charge; both years carry pension “earnings, interest and deferrals” expense. Pension is a non-operating headwind, not a tailwind. FY24 GAAP was depressed by ~$69M of one-timers (settlement + $16.9M impairment) — meaning the optical FY24→FY25 EPS jump ($3.71→$7.42) is exaggerated by a low base. FY25 is essentially clean.
• Cash converts. CFO/net-income ran 0.26x (FY23, weak — working-capital build) → 1.47x (FY24) → 1.17x (FY25, CFO $440M vs NI $376M). SBC is modest ($22.8M, 0.8% of sales); deferred tax small. The genuine caveats are (a) the surcharge timing lag (quarter-level noise), (b) a ~160-day cash-conversion cycle that makes working capital absorb incremental cash during growth (so FCF lags EPS), and © an old, ~61%-depreciated asset base that flatters ROIC/ROA versus replacement cost.

Free cash flow. FCF was negative in the downturn (−$85M FY22, −$68M FY23) as working capital consumed cash, then turned: $178M (FY24) → $286M (FY25) → FY26 guided to “at least $350M” (nine-month operating cash flow already ~2x the prior-year period). Capex is rising for Athens (FY26 ~$260M, of which ~$175–185M is the project), but the program is internally funded.

Balance sheet — strong and de-risked. Net debt/EBITDA fell from a 5.05x peak (FY22) to 0.57x (FY25) — entirely via EBITDA recovery, with absolute debt flat ~$703M. The capital structure is simple (two senior unsecured notes; per the FY25 10-K, $400M @6.375% due 2028 and $300M @7.625% due 2030, subsequently refinanced in FY26 with maturities extended to 2034 at a lower rate), with the revolver upsized from $350M to $500M and undrawn, and no near-term maturities. Combined pension/OPEB underfunding is ~$178M (<1% of market cap), improving and actively de-risked. The balance sheet comfortably funds Athens plus buybacks without new debt.

Returns. ROE 20.6%, ROIC 17.2%, ROA 11.1% (FY25) — all moot-consistent at the cycle peak, though flattered by the depreciated asset base. The honest read: these are near-peak expressions of a structurally-improved-but-still-cyclical business.

Verdict (Financial Quality): Economics genuinely improve with scale — and the reported numbers are high-quality, if anything understated. The ex-surcharge margins are the right metric and are real; CFO exceeds net income; the balance sheet is under-levered. The caveats (surcharge lag, long cash cycle, depreciated-asset flatter, near-peak returns) are about durability and normalization, not accrual integrity.

7. Capital Allocation

Sequencing — disciplined and correct. Management ran the playbook in the right order: (1) repair the balance sheet through the downturn; (2) protect, not hike, the dividend; (3) restart a modest buyback only after net leverage fell below 1x; (4) commit excess FCF to a high-return organic expansion rather than M&A.

Dividend. Quarterly $0.20 ($0.80/yr) — held flat through the COVID collapse (never cut; paid continuously since 1906). Payout collapsed from ~70% (FY23) to ~11% (FY25) as earnings recovered. Conservative and credible, but the dividend has not been raised in roughly a decade, leaving an atypically low payout — return of capital has shifted to buybacks.

Buyback. None in FY22–FY24. A $400M authorization (July 2024, no expiry) was restarted only after deleveraging; $101.9M repurchased in FY25 (~$177 average), with the pace stepping up in FY26 ($49.1M Q1 + $32.1M Q2 + $133.9M Q3 = $235.8M cumulative through Q3 FY26). Management frames buybacks as primarily “dilution-offset.” The $177 FY25 average looks shrewd against today’s ~$586 — but the program is small relative to FCF, and buying back stock at a 96th-percentile own-history multiple in FY26 is far less obviously value-accretive than the FY25 purchases were.

M&A — dormant, by design. No acquisitions since 2018. The prior additive-manufacturing deals (Puris 2017, LPW 2018) destroyed value and were impaired; Latrobe (2012, ~$558M) had a mixed reception. Correctly, management has done no deals during the recovery and signals no appetite — growth is organic/brownfield.

The central capital decision — Athens. The $400M Athens, AL brownfield expansion (announced Feb-2025) adds premium VIM melt (~9,000 tons, ~7% of SAO capacity), targeted to add ~$150M incremental operating income from ~FY2028 — a ~37% pre-tax incremental return in a capacity-constrained, oligopolistic niche where capacity is the binding industry constraint. Through a Greenwald/Marathon lens this is exactly the right place to reinvest: high-incremental-ROIC capacity behind irreplaceable assets and qualification barriers, deliberately sized too small to break industry pricing. The risks are execution/timing and that the aero up-cycle peaks before the FY28 ramp.

Incentives (FY25 proxy). CEO Tony Thene (since 2015, ex-Alcoa CFO) earned $8.4M (80:1 pay ratio) and holds ~595,628 shares (~$340M+). The annual incentive is Adjusted Operating Income + Adjusted FCF + Safety (230% attainment FY25); the long-term plan is Adjusted ROIC + Adjusted EBITDA + Sustainability + relative-TSR modifier. Rewarding Adjusted FCF and Adjusted ROIC explicitly is rare and good for a capital-intensive cyclical — it pushes against empire-building. One watch item: the annual-bonus maximum was ratcheted to 250% in the boom year.

Verdict (Capital Allocation): Intelligent and improving through the cycle. Disciplined sequencing, high-return self-funded reinvestment, no value-destructive M&A during the recovery, and well-aligned incentives. Blemishes: the impaired additive acquisitions (pre-2019), a dividend frozen so long the payout is unusually low, and the question of buying back stock at a record multiple.

8. Changes and Headwinds — Last Two Years

Strategic and leadership changes (thesis-relevant).

• CEO transition (announced Feb-2026): President & COO Brian Malloy (who ran SAO) becomes President & CEO effective July 1, 2026; Tony Thene moves to Executive Chairman. Malloy is an operations insider — the appointment telegraphs continuity of the margin/mix playbook, not a strategic break. The board expanded 11→12 directors to seat him.
• Feb-2025 investor update: set the FY27 $765–800M operating-income target and announced the $400M Athens expansion — the strategic anchor of the multi-year story.
• Track record of beating targets: the May-2023 investor-day goal of $460–500M operating income (a doubling of the pre-COVID high) was exceeded two years early (FY25 $525M adjusted OI).
• Guidance reset coming: management has repeatedly called the FY27 target “outdated” and will reset it upward plus add longer-term annual guidance on the Q4 FY26 call (~late July 2026) — the key near-term catalyst.
• Carpenter Additive / PEP restructuring turned the additive business into a small profit “bright spot”; Dynamet remains the medical-destock drag.
• Debt refinancing (FY26): notes extended to 2034 at a lower rate; revolver upsized to $500M; FCF guidance raised to “at least $350M.”

Headwinds and bear points.

• Valuation is the dominant bear point — richest-ever own-history multiple, doubled in a year, little room for an aero-cycle disappointment.
• Cyclicality / “bullwhip.” Management itself describes the aero nickel market as historically very cyclical — supply chain gets ahead, destocks, then scrambles. Carpenter is riding the up-leg; the same dynamic cuts the other way.
• OEM build-rate / engine-durability risk. The thesis leans on the 737 MAX (42→47→52) and A320neo ramps actually happening; GTF/LEAP durability issues, Boeing execution, or a rate stall would slow ordering. Management’s hedge: high aftermarket MRO content supports material pull even if new-build slips, and Carpenter is spread across platforms/OEMs.
• Customer concentration / channel disconnect. A&D 62%+ flows through GE/RTX-Pratt/Rolls/Safran/Boeing/Airbus supply chains. Notably, sell-side “channel checks” have repeatedly come in weaker than Carpenter’s reported results — an unresolved external-validation tension (management tells investors to follow Carpenter’s own data).
• Medical/PEP forecasting has been “elusive,” a minor credibility ding.
• Surcharge lag, defense lumpiness, IGT order timing make any single quarter noisy.
• Titanium structural weakness is explicitly “no material impact” (Carpenter does not melt Ti or make large Ti structural forgings — only Ti fasteners), and Western Ti tightness post-VSMPO is, if anything, a tailwind.

Verdict (Changes/Headwinds): Net STRENGTHEN the thesis — with valuation/cyclicality the offsetting risk. Record-and-rising earnings, expanding margins, a consolidated supply-constrained oligopoly, accretive capacity coming, a continuity CEO transition, and a clean balance sheet all reinforce the operating story. The four things that most matter going forward: (1) the upcoming FY27 reset + new long-term targets; (2) whether the aero-structural “bullwhip” volume actually lands; (3) Athens delivering +$150M OI from FY28 on time; (4) valuation/cyclicality — at a ~97th-percentile multiple there is little room for a stumble, and the channel-check disconnect remains unverified.

9. Risk Analysis (Risk Matrix)

The dominant risks cluster on the price/cycle axis (valuation de-rating, cycle roll, margin mean-reversion, momentum unwind), not on the business-integrity axis (accounting, balance sheet, governance are clean). This is a quality-cyclical-at-a-rich-price risk profile.

10. Valuation Discussion (Embedded Expectations)

No price target, no recommendation. This section frames what the price embeds and the scenario distribution.

Current multiples (reconciled). At $586.47, ~$29.0B market cap and ~$29.4B EV: P/E ~62x trailing ($9.51 TTM EPS), ~56x on FY26E $10.43, ~49x on FY27E ~$11.97; EV/EBITDA ~42x trailing ($700M); P/B ~12.5x (BVPS $46.68); P/S ~9.5x. The single highest-signal datum is the own-history percentile: P/E 90.4th, P/B 99.98th, P/S 99.98th, composite 96.8th — Carpenter has never been this expensive versus itself on book and sales.

Peer context. Carpenter at ~42x EV/EBITDA trades above ATI (~28–32x; forward P/E ~37.6x) and roughly in line with Howmet (~42x) — but Howmet is a ~25%-operating-margin, ~21%-ROIC, recurring-aftermarket-rich franchise with a qualification moat closer to TransDigm, whereas Carpenter is a higher-beta, more cyclical melt-and-mill producer whose margins are at a record and whose ROIC was negative as recently as FY21–22. Carpenter earning an above-ATI multiple and a Howmet-like EV/EBITDA is the crux of the debate. On forward P/E, Carpenter (~49x FY27E) sits between ATI (~37.6x) and Howmet (~44–52x).

What $586 requires (reverse / embedded expectations). The danger in a cyclical-at-peak is double-counting — a peak multiple on peak earnings. To support ~$29.4B EV at a “normal” cyclical exit of ~14–16x EV/EBITDA, the market must underwrite mid-cycle EBITDA of roughly $1.85–2.1B — about 2.6–3.0x the current ~$700M. Even crediting Athens (+~$150M OI from FY28) and a full up-cycle to ~$1.0B EBITDA by FY28, $29.4B EV implies a sustained ~29x EV/EBITDA on FY28 numbers. On earnings, $586 / FY28E consensus $14.27 ≈ 41x two years out; for a 9–10% forward return at a terminal 22–25x P/E (still a cyclical premium), EPS would need to roughly double from FY26E $10.43 toward ~$20–23 — well above the $14.27 consensus. The price is not discounting mean-reversion; it is extrapolating today’s elevated margins and a premium multiple years forward. The embedded expectation is “secular compounder,” not “cyclical at peak.”

Scenario distribution (illustrative of embedded expectations; not a price target).

• Bear (A&D cycle rolls / destock / margins mean-revert): EBITDA normalizes toward ~$520–580M, multiple de-rates toward own-history-mid / cyclical-peer ~12–15x → EV ~$7–9B — a multiple-and-earnings double-hit, the classic cyclical air-pocket (the −85% lifetime drawdown is the precedent for how violent these de-rates are).
• Base (up-cycle continues, Athens ramps, no recession): FY28 EBITDA ~$0.9–1.05B (consensus-consistent, EPS ~$14); multiple compresses from 42x toward a still-rich ~18–24x → EV ~$18–25B — implied value below to around today, i.e., low/negative annualized return even if operations stay excellent, because the multiple does the damage.
• Bull (clean multi-year ramp + Athens fully accretive + mix-up + margins push higher): FY28 EBITDA ~$1.1–1.3B; market keeps awarding a 30–35x Howmet-like multiple → EV ~$33–45B — modest upside, but requires a 96th-percentile multiple to hold through a doubling of the earnings base.

Key observation: the distribution is negatively skewed — limited upside, fat left tail. The risk is the multiple, not (only) the franchise. The one variable that could rehabilitate the math is true mid-cycle EBITDA: if the A&D step-change is structural (mid-cycle ~$800–900M, not ~$600M), the reverse-math is materially less stretched — which is precisely the open question on which the entire valuation hinges.

Most-appropriate metrics: EV/EBITDA and forward P/E, both on the ex-surcharge basis (surcharge distorts revenue/margin optics; P/B is distorted by the heavy asset base). The most useful single fact remains the 96.8th own-history composite percentile.

11. Variant Perception

Consensus belief. Carpenter is a structural beneficiary of a multi-year aerospace & defense supercycle — a quality specialty-alloy franchise with high barriers (melt capital intensity, qualification, surcharge pass-through pricing power), record-and-rising margins, ROIC inflecting from negative to ~17%, Athens adding capacity into a supply-constrained market, and a clean multi-year earnings ramp ($7.42 FY25 → ~$10.43 FY26E → ~$12 FY27E → ~$14 FY28E). The sell-side is positive; the market pays ~50–62x earnings / ~42x EV/EBITDA because it underwrites this as durable secular growth.

Strongest bull case. (1) A&D demand is structurally supply-constrained — engine OE ramp + an early-innings spares supercycle + defense + medical + energy mix-up — and Carpenter sells qualified, sole/limited-source alloys with pricing power. (2) The margin/ROIC turn is structural (re-engineered higher-value mix, exit of commodity grades), so “peak” earnings are actually a new durable baseline. (3) Athens adds ~$150M OI from FY28 into a market that absorbs it at full price. (4) Operating leverage + buyback + further mix-up drive EPS toward $14–18 by FY28–29, making today’s multiple look reasonable in hindsight.

Strongest bear case. (1) Valuation: richest-ever own-history multiple; the entire bull case is the base case in the price — no margin of safety. (2) Cyclical priced as secular: ROIC was negative FY21–22; −85% lifetime drawdown; margins and multiple are both peak — a textbook double-count. (3) Negative asymmetry: even the base case caps/erases the return via multiple normalization; a cycle wobble is a 40–60%+ drawdown. (4) Crowded, high-beta, momentum-funded ownership — vulnerable to a regime unwind. (5) Surcharge optics flatter reported revenue, and the “real” ex-surcharge economics, while good, do not obviously justify a Howmet-like multiple for a more cyclical, lower-aftermarket business.

The assumptions that matter most. (A) Is current ~$700M EBITDA / ~22% margin a peak or a new durable baseline? (B) How long and how durable is the A&D up-cycle — early-innings or mid-to-late? © Will the ~42x / 96.8th-percentile multiple hold, or compress toward 14–24x? (D) Is the margin/ROIC turn structural (sticky) or cyclical (mean-reverting)? (E) Does Athens arrive into a tight market (accretive) or a softening one (over-capacity)?

What would falsify each side. Falsify the bull: a Boeing/Airbus build-rate cut or engine-durability shock that breaks the volume ramp; an A&D destock; margins rolling toward FY24 levels; or simply a de-rate toward ATI’s ~28–32x. Falsify the bear: continued guidance raises with ex-surcharge margins pushing past ~24%, FY27–28 EPS confirming/beating ~$12–14, ROIC sustaining past 17%, Athens ramping accretively into a tight market, and the multiple holding at 40x+.

Factor-positioning input (from the tape). In factor space Carpenter is a high-beta aerospace-momentum vehicle (market beta 1.43–1.50, A&D industry loading 0.78 dominant, infrastructure 0.55, negative liquidity loading), with an exceptional recent risk-adjusted record (+134% one-year return, Sharpe 2.74) — a working momentum trade, not a falling knife today. But the −85% lifetime max drawdown and the negative-ROIC FY21–22 history argue this is, at root, a cyclical whose current low-realized-vol / high-Sharpe profile is a cycle-peak artifact. A crowded, high-beta, momentum-funded name at a 96.8th-percentile multiple is the textbook set-up for a sharp unwind if the A&D factor rolls — the very scenario the low recent drawdown lulls investors into ignoring.

Crux. Both sides largely agree the business is good and improving. The disagreement is entirely about (1) whether current economics are peak or baseline, and (2) whether a 96.8th-percentile own-history multiple can persist. The single most important variable for forward returns from $586 is not Carpenter’s near-term operating performance (likely good) but whether a record cyclical multiple on record cyclical margins can hold. That is the whole question.

12. Fact vs. Interpretation Table

13. Open Questions

1. True mid-cycle EBITDA — ~$600M (historic-cyclical) vs ~$800–900M (if the A&D step-change is structural)? Drives the entire reverse-math.
2. Peak vs. baseline margins — is ~22% ex-surcharge operating margin / 35.6% SAO a near-peak or a durable floor? The FY27 reset + long-term targets (Q4 FY26 call) will inform this.
3. LTA mix and pricing — ~40% of total / ~60–65% of aero revenue under LTA; how much repricing runway remains after two post-COVID renewal rounds?
4. Maintenance vs. growth capex split — D&A ~$139M is a rough maintenance proxy; need the split to model normalized FCF as Athens spend rolls off.
5. Surcharge-lag direction into FY26–27 — tailwind or headwind to reported margin given nickel/cobalt moves?
6. Athens demand sensitivity — does the +$150M OI target assume the up-cycle holds through the FY28 ramp? Downside if build rates disappoint.
7. Channel-check disconnect — why have sell-side checks repeatedly read weaker than reported results? Resolution either validates differentiation or flags a forward air-pocket.
8. PEP/Carpenter Additive — fixed, divestable, or a continuing drag? Medical-distribution destock trajectory.
9. Legacy environmental (Reading PA) and Dynamet CBA (expired Aug-2025) — verify reserves and labor status from 10-K legal proceedings.

14. What Must Be True

Bull case — what must be true. The A&D up-cycle is early-innings, not late; the aero-structural “bullwhip” volume actually lands on top of the engine/spares ramp; ex-surcharge margins prove structural (durably past ~22–24%, not a utilization-peak artifact); Athens ramps accretively into a still-tight market; EPS marches toward $14–18 by FY28–29; and the market continues to award a 35–40x+ EV/EBITDA / 96th-percentile multiple.

• Falsification test: a Boeing/Airbus build-rate cut, an A&D supply-chain destock, ex-surcharge SAO margins rolling back toward ~29–30%, or a multiple de-rate toward ATI’s ~28–32x EV/EBITDA — any one breaks the “pay-up-for-secular-quality” thesis. Concretely: if FY27 ex-surcharge SAO margin prints below ~32% or the multiple compresses >25% while earnings rise, the bull thesis is failing.

Bear case — what must be true. Current ~$700M EBITDA / ~22% margin is a cyclical peak; the recent price/volume gains are late-cycle; the multiple is unsustainable and mean-reverts toward own-history-mid (14–24x); a cycle wobble (build-rate cut, destock, recession) compresses earnings and multiple together for a 40–60% drawdown.

• Falsification test: continued guidance raises with ex-surcharge margins pushing past ~24%, FY27–28 EPS confirming/beating ~$12–14, ROIC sustaining/rising past 17%, Athens ramping accretively, and the multiple holding at 40x+ — which would prove the premium earned and the business structurally re-rated (Howmet-style). Concretely: if the FY27 reset lands materially above $800M OI with margins still expanding and the stock holds its multiple, the bear thesis is failing.

The two tests are mirror images, which is the point: this is not a debate about whether the business is good (both sides concede it is) but about whether peak economics at a peak multiple can persist. Watch the FY27 guidance reset (Q4 FY26 call, ~late July 2026) and the ex-surcharge SAO margin trajectory as the cleanest near-term arbiters.

15. Source Appendix

See Appendix B for the full citation list. Primary sources: Carpenter Technology FY2025 Form 10-K (filed 2025-08-12), FY2025 DEF 14A (filed 2025-09-12), FY26 Q1–Q3 and FY25 Q4 earnings-call transcripts, the Feb-2025 investor update deck, FY24–FY26 8-K corpus, and SEC Form 3/4 insider filings (CIK 0000017843). Quantitative data: ROIC.ai (statements, ratios, enterprise value), AZI (price history, valuation_index own-history percentiles), FactorsToday (factor loadings, leaderboard). All non-obvious facts are cited with source and access date in the source appendix.

APPENDIX A — Standard Diligence Questionnaire

Carpenter Technology Corporation (NYSE: CRS) — as of 2026-06-19

Supplemental to the research memo. Fact / Interpretation / Assumption labels applied where it matters.

General

What thoughtful questions have other investors asked about this company? The recurring debates: (1) Is current ~$700M EBITDA / ~22% ex-surcharge operating margin a cyclical peak or a new structural baseline? (2) Can a 96.8th-percentile own-history multiple (~42x EV/EBITDA, ~62x P/E) persist? (3) Why have sell-side channel checks repeatedly read weaker than Carpenter’s reported results — differentiation, or a forward air-pocket? (4) How much LTA repricing runway remains after two post-COVID renewal rounds? (5) Will the aero-structural “bullwhip” volume actually land on top of the engine/spares ramp?

Cyclicality & Earnings Nature

Cyclical high or low? Interpretation: Near a cyclical high on margins and multiple. ROIC went from negative (FY21–22) to 17.2% (FY25); SAO ex-surcharge margin hit a record 35.6%; −85% lifetime max drawdown confirms deep cyclicality. The volume leg (aero-structural restock) may be earlier-cycle than margins. External environment or internal actions? Both. Internal: mix shift to A&D >60%, exit of commodity grades, ~80% utilization, repricing post-COVID LTAs. External: aerospace build-rate recovery, engine spares cycle, IGT/data-center demand, Western Ti tightness post-VSMPO sanctions. How stable are revenues? Cyclical-recurring. ~40% of total / ~60–65% of aero revenue under multi-year LTAs, but tied to OEM build rates and aftermarket cadence; revenue fell ~32% peak-to-trough in COVID (FY20 $2.18B → FY21 $1.48B). Surcharge pass-through (~18% of revenue) adds reported-revenue volatility unrelated to volume. Outlook for products/services? Strong multi-year demand stack: narrowbody engine ramp, early-innings spares supercycle, defense, IGT/data-center power, medical recovery. How big is the market, growing or shrinking? Premium aerospace high-performance-alloy market ~5.5% CAGR; structurally constrained supply. Domestic and international (US/Europe/Asia). Growing.

Business Quality & Competitive Moat

Industry more or less competitive? Less — consolidating: Haynes→Acerinox (Nov-2024), Universal Stainless→Aperam (Jan-2025) removed two independents; public pure-plays now just CRS + ATI. How profitable (ROIC, ROE)? FY25 ROIC 17.2%, ROE 20.6%, ROA 11.1% — near cycle-peak, flattered by a ~61%-depreciated asset base. How profitable is the industry; competitors; barriers? Premium tier highly profitable for the <10 qualified melters; barriers are AMS qualification + capital-intensive VIM/VAR/ESR vacuum melt + multi-year requalification. Commodity tier is low-margin and fragmented (CRS exits it). Easily understood? Yes — a premium specialty-alloy melter selling qualified mill products into aerospace/medical/energy. Undermined by foreign low-cost labor? No — qualification, capital intensity and flight-safety regulation block low-cost substitution; the binding input is metallurgy/quals, not labor cost. (Russian VSMPO Ti is sanctioned out, a tailwind.) Do brands matter? Specifications and qualification incumbency matter more than consumer brand; “Carpenter” carries real metallurgical reputation/trust among OEMs. Nature of competition? On qualification incumbency, capacity availability, metallurgy/yield and delivery — not primarily price (surcharge handles metal cost; LTAs + quals limit price wars). Switching costs? High — requalifying a second source on a flight-critical grade costs years plus safety/regulatory risk; produces sole/limited-source positions.

Financial Condition & Balance Sheet

Assets not fully on the balance sheet? Interpretation: Yes — the qualification incumbency / customer relationships and the irreplaceable, deeply-depreciated melt asset base (gross PP&E $3.5B vs net $1.36B) are worth far more than book; reported ROIC/ROA are flattered by low net book value. Off-balance-sheet liabilities? Combined pension/OPEB underfunding ~$178M (improving, actively de-risked); operating leases modest; legacy environmental (Reading PA) — verify reserves in 10-K legal proceedings. How conservative is the accounting? Conservative/high-quality — surcharge dilutes (does not flatter) margins; CFO > net income (1.17x FY25); SBC modest (0.8% of sales); FY24 was depressed ~$69M by a pension settlement + impairment (so FY24→FY25 growth is exaggerated by a low base, not flattered). How CapEx-hungry? Moderately — maintenance capex ≈ D&A ~$139M; currently elevated for the $400M Athens growth project (FY26 capex ~$260M). Long ~160-day cash-conversion cycle means working capital absorbs incremental cash during growth.

Capital Allocation & Management

How much FCF; how used; philosophy? FCF $286M (FY25), guided “at least $350M” FY26. Philosophy: fund accretive growth (Athens) + return cash (buyback + dividend); balance sheet first. Disciplined sequencing. Significant acquisitions recently? None since 2018. Prior additive deals (Puris 2017, LPW 2018) were value-destructive and impaired; no M&A appetite signaled. Buying back shares? Yes — $400M authorization (July 2024); $101.9M FY25 (~$177 avg), $235.8M cumulative through Q3 FY26 (pace stepping up to $133.9M in Q3). Framed as “dilution-offset.” Issuing large amounts of stock to insiders? No — SBC modest ($22.8M); share count roughly flat ~49.7M; equity comp monetized via vesting, not large new issuance. Compensation policy? CEO Thene $8.4M FY25 (80:1 ratio); annual bonus = Adj OI + Adj FCF + Safety; LTI = Adj ROIC + Adj EBITDA + Sustainability + relative-TSR modifier. Well-aligned (FCF + ROIC explicit). Watch: bonus max ratcheted to 250% in the boom year. Motivations of management? Strong ownership alignment (CEO ~595K shares, ~$340M+); credible through-cycle stewardship (protected dividend, repaired balance sheet, beat targets early). CEO transition to Malloy (ops insider) July 2026 = continuity.

Valuation & Market Data

ADR, MLP, or K-1? No — US C-corp common stock, NYSE-listed, standard 1099 dividend. Dividend policy? $0.20/quarter ($0.80/yr), flat ~a decade, never cut (paid since 1906); payout ~11% (atypically low after the earnings recovery). Yield ~0.14%. How profitable? Very, at the cycle peak — see ROIC/ROE above; ex-surcharge operating margin ~22% (SAO ~28.6%, record quarter 35.6%). Net income diverging from cash from operations? Converging favorably — CFO $440M > NI $376M (FY25); was weak in FY23 (0.26x) on working-capital build during the ramp.

Risks & Downside

What would cause the stock to decline? Multiple de-rating (96.8th-percentile starting point); an aerospace cycle roll / build-rate cut / A&D destock; margin mean-reversion; a momentum/factor unwind (beta 1.5, A&D loading 0.78); recession. Earnings and multiple can fall together. Risk of catastrophic loss? Interpretation: Low at the business level (clean balance sheet, 0.57x net leverage, no near-term maturities, real moat) but a 40–60%+ drawdown is plausible on a cycle roll from a record multiple — the −85% lifetime max drawdown is the historical precedent. Chance of total loss? Very low — no solvency, fraud, or going-concern risk; under-levered, cash-generative, durable qualified franchise.

Recent News & Events

Business environment changed recently? Yes, favorably for operations: consolidating supply side (Haynes/Universal acquired), accelerating A&D bookings, IGT/data-center demand, repeated guidance raises (FY26 OI to $700–705M). The change against the investor is valuation — the stock doubled in a year to a record multiple. Significant acquisitions? None by CRS; the relevant M&A is competitors being absorbed by strategics (consolidation tailwind). Change in accounting policies? None material surfaced; FY24 carried one-time pension-settlement + impairment charges (normalized out). Recent changes — markets, facilities, management? $400M Athens AL brownfield melt expansion (online FY28); debt refinanced to 2034, revolver upsized to $500M (FY26); CEO transition to Brian Malloy effective July 1, 2026 (Thene → Executive Chairman); FY27 guidance reset + new long-term targets coming on the Q4 FY26 call (~late July 2026).

APPENDIX B — Source Appendix

Carpenter Technology Corporation (NYSE: CRS) — research as of 2026-06-19

Primary sources first. Facts in the memo trace to these; quantitative figures are reconciled to filings. Access date 2026-06-19 unless noted.

A. Primary — SEC filings (CIK 0000017843)

1. Form 10-K, FY2025 (filed 2025-08-12; crs-20250630.htm) — business description, two-segment structure, end-market mix, surcharge mechanics, competition & barriers (Items 1/1A), segment operating-income bridge (Note 20), net sales excluding surcharge & adjusted margins (MDA), Other expense net incl. $51.9M FY24 pension settlement, financing arrangements (senior notes, revolver, maturities), pension/OPEB funded status, PP&E, share-repurchase program (Note 13), dividend history. Mirrored locally: output/CRS/sources/10-K/2025-08-12_crs-20250630.htm.
2. Form 10-K, FY2021–FY2024 — multi-year margin/ROIC/EPS history; COVID-trough economics. output/CRS/sources/10-K/.
3. Form 10-Q, FY2026 Q1–Q3 — quarterly revenue, segment margins, per-share data, capex, buyback pace. output/CRS/sources/10-Q/.
4. DEF 14A (proxy), FY2025 (filed 2025-09-12; crs-20250911.htm) — CEO comp $8,416,878 & 80:1 pay ratio; annual incentive metrics (Adj OI + Adj FCF + Safety, 230% attainment, 250% max); LTI metrics (Adj ROIC + Adj EBITDA + Sustainability + relative-TSR modifier); ownership guidelines. output/CRS/sources/DEF_14A/.
5. Form 8-K corpus, 2024-07 → 2026-04 — quarterly earnings; $400M buyback authorization (July 2024); 2025-02-18 Investor Update (Athens $400M expansion + FY27 target); 2026-02-17 CEO transition (d109893d8k.htm; Malloy → CEO 7/1/26, Thene → Executive Chairman); 2026-04 board expansion to 12. output/CRS/sources/8-K/.
6. Form 3/4/5 insider filings, 2025-05 → 2026-06 (241 Form 4s sampled) — zero open-market purchases (code P); routine RSU vesting (code F), near-expiry option exercises (M→S), small director sales; CEO Thene ~595,628 shares post-vesting. output/CRS/sources/4/, /3/.

B. Primary — earnings-call transcripts (ROIC.ai MCP)

7. FY2026 Q3 call, 2026-04-29 — record OI $186.5M; SAO 35.6% ex-surcharge margin (17th consecutive quarter of expansion); FY26 OI guide raised to $700–705M (“at least +33%”); FY27 target “outdated… update next call” + longer-term guidance; Athens on track; buyback $133.9M (cumulative $235.8M); FCF “at least $350M.”
8. FY2026 Q2 call, 2026-01-29 — FY26 guide $680–700M; supply/demand & nickel-superalloy supply section; 3 aero LTAs at “significant” price increases; debt refi to 2034, revolver →$500M; aero-structural “off the sidelines.”
9. FY2026 Q1 call, 2025-10-23 — 5 aero LTAs; bookings +23% seq; FAA 38→42 approval; titanium “no material impact”; transactional pricing ≥ LTA.
10. FY2025 Q4 call, 2025-07-31 — FY25 OI $525.4M; FY26 guide $660–700M set; FY27 $765–800M; PowerGen +100%+ YoY; Paris Air Show LTAs.

C. Primary — investor materials

11. Carpenter Investor Update deck, 2025-02-18 — FY27 $765–800M operating-income target; $400M Athens VIM brownfield, ~9,000 tons (~7% over FY19), +$150M OI from ~FY28. https://s21.q4cdn.com/563831349/files/doc_presentations/2025/02/1/CRS-Investor-Update-Event-Presentation-250218.pdf
12. CEO transition release — https://ir.carpentertechnology.com/news-events/news/news-details/2026/... (Malloy → CEO 7/1/26).
13. Athens expansion coverage — AIST https://www.aist.org/carpenter-technology-to-expand-premium-melt-capacity; SEDC https://www.sedc.org/news/....

D. Quantitative data services

14. ROIC.ai MCP — income statement, balance sheet, cash flow, profitability/credit/liquidity ratios, enterprise value, valuation multiples, per-share data (FY20–FY25 annual + FY24–FY26 quarterly). EV/EBITDA, ROIC 17.2%, net debt/EBITDA 0.57x. Third-party aggregated; reconciled to filings.
15. AZI — 5-year daily price CSV (closes, EMAs; COVID low $13.69 2020-03-23, 5yr low $24.59 2022-07-11, 52wk low $230.34 2025-09-25, ATH $586.01 2026-06-18); valuation_index own-history percentiles (P/E 90.4, P/B 99.98, P/S 99.98, composite 96.8). News feed (sparse for CRS).
16. FactorsToday — stock-loadings (market beta 1.43, Aerospace & Defense 0.78, Infrastructure 0.55, Liquidity −0.78); leaderboard (y1 +134% return, Sharpe 2.74, lifetime max DD −85%); related-stocks (ATI 0.96 similarity, HWM, BWXT). Third-party statistical estimates.

E. Peer / industry cross-read

17. Howmet Aerospace (HWM) report, 2026-06-13 — value-chain framing (HWM caster node above CRS melt node), moat/industry comparison, EV/EBITDA ~42x, own-history composite 95.2. output/HWM_2026-06-13_full_report.md.
18. ATI Inc. comps — valueinvesting.io/ATI (EV/EBITDA 32.2x), stockanalysis.com/stocks/ati (28.0x, fwd P/E 37.64x), ATI 2026 guide adj EBITDA $975M–$1.025B, A&D ~68%→70%.
19. Consensus EPS — Zacks via dailypolitical.com (2026-05-21), themarketsdaily.com (2026-06-05), investing.com CRS consensus: FY26E ~$10.43, FY27E ~$11.93–12.01, FY28E ~$14.27.
20. Industry / competitor facts — Acerinox/SEC 8-K (Haynes acquisition closed 2024-11-21, ~$970M EV); Aperam/SEC 8-K (Universal Stainless closed 2025-01-23, ~$537M EV); AIST/machinedesign/alloys.com (VIM/VAR/ESR mandatory AMS route); Persistence/Spherical market research (aero HPA ~5.5% CAGR); ForcedAlpha CRS playbook (LEAP+GTF ~1,300–1,700 t/yr incremental superalloy by 2028).

All non-obvious facts are supported by a primary source with URL and access date. Where ROIC/AZI/FactorsToday (third-party aggregated data) and a primary filing differ on a material number, the filing governs.