IES Holdings, Inc. (NASDAQ: IESC) — A Brilliant Data-Center Compounder at Peak Cycle, Peak Multiple, and Peak Insider Selling

Independent equity research. Report date: 2026-06-21. Fiscal year ends September 30. All figures from SEC filings (FY2025 10-K filed 2025-11-21; Q2-FY2026 10-Q filed 2026-05-01; DEF 14A filed 2026-01-07), company disclosures, and public market data unless noted. Price as of 2026-06-18 close ($712.39).

⚡ Claude’s Take

This block is the author’s own subjective opinion and general information only — not investment advice. The analysis that follows it takes no position and carries no price target; any opinion and valuation zone are confined to this block.

Verdict: AVOID-here / HOLD-don’t-chase. Not-a-short. Accumulate only on a major cyclical washout. Fair-value zone ~$400–550; the entry I’d actually want is ~$350–450. Medium-high conviction that risk/reward is poor here, lower conviction on timing. At ~$712 you are paying ~38x GAAP / ~45x the company’s own adjusted EPS / ~28x EV/EBITDA — the 98.9th percentile of IES’s own decade on every metric — for a cyclical contractor at peak margins.

This is a genuinely brilliant business at a genuinely terrible entry point, and the framing is late-stage, extreme-momentum quality-cyclical at a triple peak — cycle, multiple, and insider selling. Give IES every bit of its due: it is a superbly run, Tontine-controlled, decentralized “Constellation-for-the-trades” that has compounded ~28x off its 2022 low by riding the AI/data-center build-out, with two segments (Communications + Infrastructure Solutions, ~49% of revenue / ~65% of segment profit) levered directly to hyperscaler capex at 15–24% margins, a near-net-cash balance sheet, and an asset-light ~34% ROIC. But every one of those superlatives is a cyclical-peak number. Operating margin went 2.6% (FY22) → 11.4% (FY25); ROIC went 9% → 34% in three years — that is the textbook signature of a cyclical peak, not a moat. Contracting is a thin-moat, low-barrier, hyper-fragmented business; IES’s only durable edge is the capital-allocation jockey (Gendell/Simmes) and an industry-wide skilled-labor bottleneck that delays, not repeals, mean reversion. Three things compound the valuation risk: (1) ~20–25% of recent pretax income is non-operating, mark-to-market securities gains (Gendell running a Tontine-style equity book through the P&L, chiefly Gulf Island Fabrication) — now largely gone as GIFI is consolidated, pulling a prop from under the multiple; (2) the ~39%-of-revenue Residential segment is a low-margin, housing-cyclical drag in a down-cycle (Q2-FY26 segment margin collapsed to ~2%); and (3) — the loudest tell — the 53%-controlling owner and every top officer have been dumping stock into the all-time high: Gendell/Tontine sold ~$118M, all discretionary (non-10b5-1), at $670–764, joined by the CEO, CFO, and a director, with zero open-market buys, ever. Comp has no ROIC governor and its metric even banks the securities gains.

Why not a short, and why not a panic: the data-center demand is real and the backlog is at records; momentum is ferocious (beta 2.07, +162% over twelve months); the float is tightly controlled; and “expensive cyclical with a labor moat” can stay expensive while the capex super-cycle runs. But buying at 38–45x into peak margins, vanishing investment-gain tailwinds, and a wholesale insider exit is paying the top dollar for the top of the cycle. Tag: “Own the jockey — but not when the jockey is the one selling.” Conviction: med-high it’s overpriced; bull-flip: a multi-year, durable hyperscaler-capex plateau plus an easing labor constraint that proves 11%+ margins are the new structural floor. Bear-flip: a data-center capex digestion or a housing-led downturn that snaps op margins back toward mid-cycle (~6–8%) while the securities-gain prop is gone — normalized EPS resets toward ~$10–13 and a 38x multiple compresses hard.

📈 Stock Price Action — Five-Year Event Map

IES is one of the great compounders of the cycle: roughly a 28-bagger from a 2022 low of $25.27 (May-2022) to an all-time high of $749.83 (Jun-12-2026), now $712.39 (~5% off the high), a ~$13.7B company. The arc is a near-pure leveraged play on the data-center build-out layered on a high-beta (2.07), low-float, controlled stock — punctuated by violent two-way swings (single-day moves of −24.6%, −19.9%, +17.9%) that are the fingerprints of a momentum vehicle, not a steady compounder. The trailing-twelve-month return is +162%.

Cycle narrative. (1) The 2022 low coincided with IES’s operating-margin trough (2.6%) — proof of how cyclical the base business is. (2–3) The 2022→2024 explosion is the data-center inflection: as hyperscaler capex surged, IES’s operating margin tripled and EPS went from $1.66 (FY22) to $10.73 (FY24), and the stock re-rated from “cheap cyclical” to “AI-infrastructure compounder.” (4 & 6) The −24.6% (Jan-2025) and −19.9% (Jan-2026) single-day drops show the other side of a beta-2.07 momentum name — when a quarter disappoints lofty expectations, the unwind is brutal. (5 & 7) Each renewed data-center beat (Q4-FY25, Q2-FY26) sent it to new highs, the Q2-FY26 print driving a melt-up to the $750 ATH. (8) The most recent wrinkle is the tell: as the stock hit its high, the controlling owner and every top officer filed Form 4s for ~$118M of discretionary open-market sales, and the stock dipped on the disclosure. (All price moves are Fact, from the AZI 5-year CSV; attributed drivers are Interpretation, cross-referenced to earnings dates, 8-Ks, and the news feed.)

1. Executive Summary

IES Holdings is a decentralized holding company of electrical, mechanical, and infrastructure-services contractors — a “Constellation/Berkshire-for-the-trades” — that has been transformed by the AI/data-center build-out. The former Integrated Electrical Services (which emerged from bankruptcy in 2006), IES is controlled (~53%) by Jeffrey Gendell/Tontine Capital and run through four autonomous segments: Communications (data-center technology infrastructure — the growth engine), Residential (single-family and multifamily electrical/HVAC/plumbing — the cyclical anchor), Infrastructure Solutions (custom-engineered power products, including generator enclosures for data centers — the highest-margin segment), and Commercial & Industrial. FY2025 (September) revenue was $3.37B (+17%), with diluted EPS of $15.16 and a near-net-cash balance sheet.

The transformation is real and spectacular — and almost entirely a cyclical peak. Revenue tripled from FY20 ($1.19B) to FY25 ($3.37B); diluted EPS went ~9x from $1.66 (FY22) to $15.16 (FY25); operating margin expanded from 2.6% (FY22) to 11.4% (FY25); ROIC went from 9% to 34%; ROE is ~47%. Two segments tied to the same hyperscaler end-market (Communications + Infrastructure Solutions) now produce ~49% of revenue and ~65% of segment profit at 15–24% margins, on record backlog. The stock has compounded ~28x off its 2022 low to a ~$13.7B all-time-high valuation.

But this is a thin-moat, cyclical contractor at a triple peak — cycle, multiple, and insider selling. Three structural cautions define the thesis. First, the moat is a cyclical-peak-returns illusion: contracting is a hyper-fragmented, low-barrier, bid-to-bid business; IES’s 9%→34% ROIC arc and 2.6%→11.4% margin arc are the signature of a capacity-constrained cyclical windfall, not a durable franchise — the only genuine edges are the capital-allocation jockey (Gendell/Simmes) and an industry-wide skilled-labor shortage that delays mean reversion. Second, ~20–25% of recent pretax income is non-operating, mark-to-market securities gains (Gendell running a Tontine-style equity book — chiefly the Gulf Island Fabrication stake — through the income statement), now largely disappearing as GIFI is consolidated; on the company’s own adjusted EPS (~$15–16, stripping these gains), the stock trades at ~45x. Third — the loudest signal — the 53%-controlling owner and every top officer have sold ~$118M of stock into the all-time high, all discretionary (non-10b5-1), at $670–764, with zero open-market buys ever, while comp carries no ROIC governor and its metric even banks the securities gains.

Net: a superb operator and capital allocator riding a genuine demand super-cycle, priced at 38–45x earnings near a record on extreme momentum (beta 2.07, +162% over twelve months), with peak margins, a vanishing investment-gain tailwind, a housing-cyclical drag, and a wholesale insider exit. The embedded expectation is a multi-year continuation of peak conditions; the disconfirming risks (data-center capex digestion, housing downturn, margin mean reversion) are exactly the ones a cyclical at peak multiples is least able to absorb.

2. Business Overview

What IES actually is. IES Holdings is a decentralized holding company of electrical, mechanical, and infrastructure-services contracting businesses. The 10-K states the model: “each of the Company’s segments manages its own day-to-day operations… Our corporate office is focused on significant capital allocation decisions, investment activities and selection of segment leadership.” The holdco overhead is tiny — ~$52.5M of corporate SG&A on $3.37B of revenue (~1.6%). The legal entity is the former Integrated Electrical Services, a 1990s electrical-contractor roll-up that over-levered and emerged from Chapter 11 in 2006; the pivotal event since is control by Jeffrey Gendell / Tontine Capital (~53% of shares). Gendell was CEO (2020–2025) and became Executive Chairman on 2025-07-01; Matt Simmes (30-year IES veteran) is President & CEO. This is a controlled company — Tontine “can control most of our affairs,” and most of its stake sits on a resale shelf (a structural overhang).

How it makes money. IES is a project-based contractor, not a recurring-revenue franchise. It designs, builds, and installs electrical/technology/mechanical systems under fixed-price and time-and-materials/unit-price contracts, recognizing most revenue on percentage-of-completion. The economic engine is labor + materials markup on project execution, with profitability driven by bidding discipline, execution, prefabrication, and procurement — not subscriptions, switching costs, or installed-base annuities. Recurring revenue is thin (only Infrastructure Solutions’ apparatus-repair MSA work and a modest maintenance book); management’s own language frames a “recurring revenue model” as an aspiration, not the current reality.

The four segments (FY2025):

The mix story the table tells is central: Residential is still the largest segment by revenue (38.7%) but the smallest by margin (8.0%) and is shrinking, while Communications and Infrastructure Solutions together are 48.6% of revenue but ~65% of segment operating income at much fatter margins (14.6% and a remarkable 23.8%). The profit center has decisively shifted to the data-center complex.

Segment detail. Communications (+46.9% YoY) is a nationwide data-center technology integrator for co-location and hyperscale customers, at record backlog — the growth and re-rating engine; Q2-FY26 revenue +34.7% YoY. Residential (−6.1%) is housing-starts-levered, declined on affordability/rate headwinds, with management guiding multifamily down in FY26; the Q2-FY26 print was worse — segment operating income collapsed to ~$6.4M (~2% margin), the cyclical soft spot. Infrastructure Solutions (+42.1%, 23.8% margin) makes custom-engineered products “such as generator enclosures used in data centers” — a second data-center engine, manufacturing-flavored, pulled by the same demand (Q2-FY26 revenue +63.6%). Commercial & Industrial (+16.2%) is lower-margin general electrical/mechanical, increasingly data-center/power-tilted.

Backlog. Total backlog reached $2,373.8M at 9/30/25, up 32.9% YoY, “primarily driven by strong demand and increased market share,” with the data-center complex dominating the growth (Communications and C&I backlog up sharply; Residential backlog fell). Caveats from the 10-K: single-family residential and much short-cycle Communications work is excluded, and ~29% of total backlog is “agreements without an enforceable obligation” (letters of intent) — real and at records, but not a contractual guarantee.

Quality-of-earnings flag (carried to the Financial Quality section). Reported EPS is inflated by non-operating investment gains — Q2-FY26 carried a $37.3M “gain on marketable securities” (vs $5.5M prior year), substantially the Gulf Island Fabrication holding remeasured to fair value. A meaningful slice of recent EPS growth is mark-to-market investment income, not contracting earnings; operating income and segment OI are the cleaner read.

Verdict: a well-run, tightly-controlled, decentralized holdco of project-based contractors transformed by the data-center build-out — a contractor, not a recurring-revenue compounder. The durable asset, if any, is the central capital-allocation engine and Tontine control, not any single business; and reported EPS is flattered by non-operating securities gains.

3. Industry Dynamics

Industry structure: vast, hyper-fragmented, low-barrier at the base. The US electrical/mechanical (MEP) and infrastructure-contracting market is a multi-hundred-billion-dollar industry of thousands of local and regional contractors with low barriers to entry — a two-truck electrical shop is trivial to start; the base-state industry is local, cyclical, price-competitive, and commoditized. At ~$3.4B revenue IES is a large player but a low-single-digit share of any national pool. The investable question — identical to the EMCOR/Comfort Systems debate — is whether the complex, large-scale, multi-trade slice (hyperscale data centers, power infrastructure, custom-engineered products) carries higher barriers than the base. The evidence says modestly yes, and rising with scope density — but it remains a thin-moat industry enjoying an extraordinary moment.

Demand — a genuine secular super-cycle, currently supply-constrained. The dominant driver across Communications, Infrastructure Solutions, and C&I is the AI/data-center build-out: hyperscaler capex at a historic peak, and modern AI data centers carry multiples of the electrical/MEP content per square foot of conventional facilities (higher power density, liquid cooling, redundant distribution, backup generation). This scope-density tailwind is a content/price lever independent of unit growth and favors sophisticated contractors. Layered on are reshoring/advanced manufacturing, grid electrification, and power-market spend. Management is explicit that growth is now labor-/capacity-limited, not demand-limited (“availability of labor and capacity could constrain the rate at which we are able to grow this business”).

The Residential counter-cycle. Cutting the other way, ~39% of IES revenue is housing-cyclical. Single-family and multifamily demand fell through FY25 on affordability/rate pressure; multifamily backlog declined; management guides multifamily down in FY26. So IES is a barbell: a booming, supply-constrained data-center complex bolted to a contracting, rate-sensitive housing business. The two partly offset (and a housing recovery is a future option if rates fall), but consolidated results blend a secular boom with a cyclical trough.

Competitive structure and the labor constraint. The scaled competitors are EMCOR (~$17B revenue), Comfort Systems USA, MYR Group, Quanta, MasTec, and Sterling — plus thousands of small shops. IES is among the smaller scaled players, more Residential- and data-center-tilted than EMCOR (which carries a ~26% recurring facilities/industrial-services ballast IES lacks). The skilled-trades shortage (journeyman electricians, pipefitters, and especially field supervisors) is the industry’s binding constraint and the crux of the bull case: you cannot conjure trained electricians with capital, so scaled firms with assembled, trained, retained workforces win the right to select and price work — exactly what the margin step-up reflects. IES runs two residential education centers and leans on prefabrication. But this edge is shared across the scaled cohort (EMCOR/FIX/MYR make the identical claim) and cyclical (in a downturn the trained workforce becomes fixed-cost deadweight, as FY22’s 2.6% margin showed).

Pricing power — rented, not owned. The current windfall (gross margin 25.5% FY25 vs 18.7% FY23; operating margin 11.4% vs 6.1%) reflects a temporary supply/demand imbalance: more complex work than the industry can staff, letting scaled firms be selective and price fixed-price risk richly. This pricing power is cyclical and capacity-driven — rented from the labor shortage, not structural.

Marathon capital-cycle read. Where is capital flooding? Into data-center construction demand: yes, at a historic peak — by capital-cycle logic the customers (hyperscalers) are over-building, and that capex eventually digests, taking IES’s record data-center backlog with it. Into MEP contracting capacity (the supply side that sets pricing): largely blocked by labor — the normal self-correction (competitors add capacity, compete margins away) is delayed because you can’t hire journeymen at will. But this is a delay in mean-reversion, not its repeal: if hyperscaler capex rolls over before the labor shortage eases, IES still de-rates, because the incremental backlog is data-center backlog. The classic capital-cycle setup: abnormal returns, generational customer over-investment, peak multiples, with the only brake a non-replicable labor constraint.

Verdict — structurally MIXED. A mediocre base industry (fragmented, low-barrier, cyclical) enjoying a genuinely excellent moment because a generational demand shock collides with a hard supply constraint, handing scaled incumbents temporary but real pricing power. In the Marathon framework the industry is mid-to-late boom: record customer capex, peak margins, peak multiples, with mean-reversion delayed (not cancelled) by the labor bottleneck. IES is more cyclically exposed than EMCOR — a ~39% housing-levered drag plus less diversification/recurring revenue — so it captures more data-center upside per revenue dollar but lacks EMCOR’s downside cushion.

4. Competitive Position

Start skeptical: contracting is a thin-moat business. By Greenwald’s tests, the base MEP-contracting industry has no genuine barrier to entry — thousands of competitors, low entry barriers, minimal switching costs, no proprietary technology, no network effects. The default verdict for a trades contractor is “no moat — run it efficiently.” The honest question is whether IES has anything that suspends mean reversion. Four candidates:

(a) Scale/density and reputation in data-center work — real but narrow and rentable. IES’s Communications segment is a nationwide data-center integrator that can staff and execute large, complex, multi-site projects a two-truck shop cannot — a partial economies-of-scale-plus-modest-captivity edge at the complex-project tier (hyperscalers prize execution certainty, schedule reliability, prefab capability, safety record; there is agency-relationship stickiness). The generator-enclosure business adds a manufacturing/engineering edge harder to replicate than field labor. But switching costs are low (each project re-bid), customers are concentrated and powerful, and the work is fixed-price competitive bidding. The edge is real at the complex tier but shallow, project-by-project, and partly rented from the current capacity shortage.

(b) The decentralized-holdco capital-allocation engine — the most defensible “moat,” and it’s a process, not a business. The strongest durable-advantage candidate is the IES operating/capital-allocation model itself — a mini-Constellation/Roper for the trades: Tontine/Gendell’s centralized capital allocation redeploying segment cash into accretive acquisitions, buybacks, and the best-returning segments; a decentralized structure retaining entrepreneurial operators; and a permanent-capital, controlled-owner orientation avoiding the over-leverage that bankrupted the predecessor. But if this is the moat, it is a management/organizational advantage — which Greenwald is explicit can be emulated and is not a structural barrier. It is real and value-creating but depends on continued disciplined execution and key people (Gendell, now 66 and Executive Chairman; Simmes). The 2006-bankruptcy history is a reminder that the same legal entity with worse discipline destroyed itself.

© Labor access in a skilled-trade shortage — the best structural argument, but shared and cyclical. The most genuinely structural edge is the assembled, trained, retained skilled workforce (10,283 employees, two education centers) in a market where labor is the binding constraint. But it is shared with every scaled player and is only valuable while demand exceeds labor supply — a delay on mean reversion, not a permanent moat.

(d) Switching costs — minimal. Each project is competitively bid; no installed-base lock-in, no proprietary platform. The thin MSA book is the only modest exception.

The ROIC test — the cyclical-peak tell. ROIC ran 16.7% (FY20) → 8.8% (FY22 trough) → 35.1% (FY24) → 33.8% (FY25); operating margin 4.8% → 2.6% → 10.4% → 11.4%. A genuine moat produces sustained 15–25%+ ROIC across a cycle. IES’s ROIC was single-digit just three years ago and is now 34% — the textbook signature of a cyclical peak, not a structural fortress. (The ~47% ROE is further flattered by an asset-light, negative-working-capital model funded by customer-billings float — ROE was already ~47–48% in FY20–21 when ROIC was only ~17%, so the high ROE is a denominator artifact, not evidence of a wider moat.)

Head-to-head. IES sits in the middle of the scaled cohort: operating margin 11.4% above EMCOR’s clean ~9% but below Comfort Systems’ ~14%; worse diversification than EMCOR (no ~26% recurring services ballast) and roughly comparable to FIX; distinguished by a ~39% housing-cyclical Residential segment (a vulnerability) and a manufacturing-flavored Infrastructure Solutions products segment (a genuine differentiator). Its most differentiated feature is the controlled-owner/decentralized capital-allocation structure — the closest thing to a non-replicable edge, and the basis of the “Constellation-for-trades” multiple argument.

Verdict — a well-run, well-allocated decentralized contracting roll-up riding a data-center super-cycle, NOT a structurally-moated franchise. The only candidate for a durable advantage is the central capital-allocation/operating model under Tontine control — a Constellation-style process edge that Greenwald classifies as superior management (emulable, key-person-dependent, not a structural barrier). The data-center scale edge is real but narrow and rented; the labor edge is real but shared and cyclical; switching costs are negligible. The 9%→34% ROIC arc is the smoking gun: this is a cyclical-peak-returns illusion dressed as a moat. The right framing is a high-quality operator and capital allocator in a thin-moat, cyclical industry — own the jockey, not a horse with a fortress around it, and respect that 34% ROIC will not persist through a data-center digestion or housing downturn.

5. Growth History and Forward Opportunities

Historical growth has been explosive — and front-loaded onto the data-center boom. Revenue compounded from $1.19B (FY20) to $3.37B (FY25), with the acceleration concentrated in FY24–25; diluted EPS went $1.97 → $3.16 → $1.66 (FY22 trough) → $5.30 → $10.73 → $15.16. The EPS path is ~9x in three years, driven by revenue growth and a tripling of operating margin (2.6% → 11.4%) and a slug of non-operating securities gains. This is not a smooth multi-decade compounding record — it is a violent operationally-levered inflection coinciding precisely with the AI-capex wave.

The forward opportunity set is genuine but concentrated and cyclical:

• Data centers (the engine) — Communications at record backlog; Infrastructure Solutions’ generator enclosures surging (+64% in Q2-FY26); C&I increasingly power/data-center-tilted. Hyperscaler capex remains the dominant driver, with scope-density (more electrical content per facility) a structural content lever. Concentration caveat: Communications and Infrastructure Solutions ride the same end-market, so “diversification” is narrower than the segment count implies, and contract scope-expansion could be pull-forward.
• Reshoring / electrification / grid — advanced manufacturing, semiconductors, EV/battery, and power-infrastructure spend feed C&I and Communications.
• Residential recovery option — ~39% of revenue is housing-cyclical and currently in a down-cycle; a rate-driven housing recovery is a future upside option (and a current drag).
• M&A — the decentralized model’s bolt-on engine (recently GIFI, Broadwind Abilene) extends capacity and adjacencies, funded by an under-levered balance sheet.

The forward reality: growth is now labor-constrained, demand is at a cyclical peak, and the recent earnings carry low-quality components. Management frames growth as capacity-limited; the Residential drag is live (Q2-FY26 segment margin ~2%); and ~20–25% of recent pretax income is non-operating securities gains that largely vanish as GIFI consolidates. So the quality of forward growth is lower than the headline trajectory: the durable part (data-center contracting) is genuine but cyclical and concentrated; the housing part is a drag; and the investment-gain part is disappearing.

Verdict — high-magnitude, low-durability growth. The data-center engine is real and the backlog is at records, but the growth is concentrated in one end-market at a cyclical-and-capex peak, dragged by a housing down-cycle, and recently flattered by non-operating gains. Magnitude is not the issue; durability and quality at this point in the cycle are.

6. Financial Quality

Margins and returns — elite, and entirely cyclical-peak. Gross margin rose from 14.7% (FY22) to 25.5% (FY25); operating margin from 2.6% to 11.4%; ROIC from 9% to 34%; ROE to ~47%. These are spectacular — and three years ago they were single-digit (margins) and ~9% (ROIC). The asset-light, negative-working-capital model (capex ~2% of sales, funded partly by customer-billings float) structurally amplifies ROE, but the direction and magnitude of the move track the data-center cycle precisely. This is peak-cycle profitability, not a through-cycle franchise return.

Quality of earnings — the central flag: non-operating securities gains. IES holds marketable equity securities as trading securities, marked through a “Gain on marketable securities” P&L line. The magnitudes are material: Q2-FY26 $37.3M (~25% of quarterly pretax income); H1-FY26 $54.2M (~20.5% of H1 pretax income), versus $7.5M in all of FY25. This is Gendell running a Tontine-style equity book — chiefly the Gulf Island Fabrication (GIFI) stake — through the operating company’s income statement. The company’s own adjusted EPS strips these out (Q2-FY26 GAAP $5.44 vs Adjusted $4.16). Two consequences: (1) the headline “~$18.85 TTM EPS” overstates the contracting earnings power — on adjusted EPS (~$15–16) the stock trades at ~45x, richer than the 38x GAAP headline; and (2) IES bought GIFI outright in January 2026 (~$152M, at $12/share), converting the holding to a consolidated, loss-making subsidiary — so this specific tailwind largely disappears going forward, removing a prop from EPS just as the multiple sits at a record. These marks are also two-way: a ~$54M H1 tailwind can become a multi-tens-of-million headwind in a down market.

Cash flow and balance sheet — clean and conservative. FY25 operating cash flow was $286M; capex $67M (~2% of sales) → FCF ~$219M, with FCF/NI ~0.94x (depressed by a $63.8M working-capital build on receivables/unbilled — a growth artifact, not a quality flag). SBC is tiny (~$13M). The balance sheet is near net-cash — minimal debt, net leverage <0.5x EBITDA — i.e., materially under-levered for a 13%-EBITDA-margin, low-capex generator (dry powder for the M&A engine, but idle capital diluting returns). Goodwill is light (tangible book is most of book value), consistent with cheap historical bolt-ons.

Verdict — clean accounting and a fortress balance sheet, but reported earnings are doubly flattered: by peak-cycle operating margins and by non-operating securities gains that are now disappearing. The honest normalized picture is materially below the $18.85 TTM headline: adjusted (ex-securities-gains) EPS is ~$15–16 at peak margins, and a mid-cycle margin normalization would pull it lower still. Read operating income, strip the securities gains, and haircut for cycle position.

7. Capital Allocation

The decentralized-holdco engine — real, but the record is short and cyclically flattered. IES runs a Danaher/Constellation-style decentralized model: autonomous segments, capital allocation centralized under Gendell, light corporate overhead (~1.6% of revenue). It throws off cash (FY25 FCF ~$219M) and shows per-share discipline. But the “mini-Constellation” narrative is only partly earned: EBITDA inflected from $81.7M (FY22) to $431.5M (FY25), mapping exactly onto the data-center boom — the “compounding machine” rests on ~3 boom years, not a through-cycle record (FY22 FCF was negative; net income was $40M just three years ago). The recent earnings power is cyclically elevated and should not be capitalized at peak as a secular compounder.

M&A — disciplined small bolt-ons, but the largest recent deal is a turnaround. Acquisition spend is steady and cheap (light goodwill = disciplined multiples): Greiner (~$67M, FY24) and numerous tuck-ins. The two most recent are bigger and lower-quality: Gulf Island Fabrication (100% of GIFI, ~$152M, closed Jan-2026, into Infrastructure Solutions) produced an operating loss in its stub quarter (a turnaround/capacity play, not an accretive bolt-on out of the gate); and Broadwind’s Abilene, TX wind-tower facility (up to ~$19.5M, May-2026) is a capacity/asset acquisition. The M&A model is genuine but the headline deals are now capacity-and-turnaround, not the high-return tuck-ins of the past.

Buybacks — mistimed, the opposite of the bull narrative. No dividend ever. A $200M repurchase program was authorized July-2024 ($168M remaining at 9/30/25). The pattern shows no counter-cyclical discipline: IES bought back the least when the stock was cheapest (FY22–23, ~$25–70) and the most when expensive (FY24–25, ~$170+); at ~$712 and ~38–45x, the muted recent open-market pace (only ~$30M of the authorization deployed in FY25; just 4,112 shares in Q2-FY26) is at least consistent with not chasing the stock at peak — but the historical timing was poor.

Balance sheet — under-levered. Minimal net debt, net leverage <0.5x EBITDA, trivial interest expense. Deployment priority is M&A first, buybacks second, no dividend. Dry powder for the Gendell engine — a positive for optionality — but idle cash plus a securities book that flatters reported ROE.

The governance/insider crux — a controlling owner and all officers selling into the high, and a comp metric with no returns governor. This is the loudest single signal in the file. Gendell/Tontine owns ~53% of the company (a controlled company, weak minority protections by design). Into the all-time high, Gendell/Tontine sold ~169,000 shares (~$118M) in May–June 2026 at $670–764 — all discretionary (non-10b5-1) open-market sales — joined by CEO Simmes, CFO McLauchlin, and a director (Cleveland), with the June-12 cluster driving the 6/17 dip. Across the multi-year corpus there are zero open-market purchases (code P) by any insider, at any price. For fairness: Gendell’s sales are ~1.5% of a very large position (paced distribution/diversification, not a panic exit), and at 53% ownership some selling is mechanically necessary for liquidity — but the timing (all at the ATH), discretionary nature, breadth (all NEOs + director), and total absence of buying make it a genuine, quantified negative. Compounding it, executive comp runs on a single metric — Adjusted Pretax Income — with no ROIC/returns/per-share/relative-TSR governor (rewards empire-building on a roll-up), and the metric does not exclude the marketable-securities gains, so the ~$54M of non-operating investment marks flowed straight into the comp base, helping push FY25 payouts to 129.2% of target. Executives are partly paid on the controlling shareholder’s stock-picking.

Verdict — above-average but flattered, with a governance asterisk. Real per-share discipline and a capital-light cash engine, undermined by mistimed buybacks, a margin-dilutive largest deal (GIFI), an under-levered sheet, and — most importantly — a controlling-shareholder insider group distributing ~$118M into the all-time high on discretionary sales with zero buys, while the comp metric banks the volatile, now-disappearing securities gains. Good operator; not an aligned, returns-governed capital-allocation machine at this moment.

8. Changes and Headwinds — Last Two Years

Strategic / corporate. (1) Leadership — Gendell moved from CEO to Executive Chairman (7/1/2025); Simmes became President & CEO. (2) Gulf Island Fabrication acquired outright (~$152M, Jan-2026), converting a trading-security stake to a consolidated subsidiary. (3) Broadwind Abilene facility asset acquisition (May-2026) — capacity for Infrastructure Solutions. (4) Steady bolt-on M&A and a $200M buyback authorization (July-2024).

Operational headwinds. (1) Residential down-cycle — ~39% of revenue, housing-starts-levered, with multifamily guided down and Q2-FY26 segment margin ~2%. (2) Data-center concentration — Communications + Infrastructure Solutions ride the same hyperscaler end-market; a capex digestion would hit both. (3) Labor/capacity constraint — caps the growth rate. (4) Vanishing securities-gain tailwind — GIFI consolidation removes the largest mark-to-market contributor.

Market / governance. (1) Extreme momentum and valuation — +162% over twelve months, ~38–45x earnings, 98.9th-percentile own-history multiples, beta 2.07. (2) Insider selling — ~$118M by Gendell/Tontine + officers into the ATH. (3) Tontine resale-shelf overhang — a controlled-company structure where the majority owner is a distributing hedge fund.

Verdict — net thesis-negative at the margin. The franchise is executing superbly into a genuine super-cycle, and the M&A optionality and backlog are real positives. But the recent changes skew cautionary: leadership transition, a lower-quality largest deal, a disappearing earnings prop, a housing drag, and a wholesale insider exit at the high. None breaks the business; together they argue the easy money has been made.

9. Risk Analysis

Catastrophic-loss assessment: low at the business level (near net-cash, asset-light, real demand), but the drawdown risk to the equity is high — a beta-2.07 momentum name at 38–45x peak earnings can lose 40–60% in a cyclical/multiple reset without the business ever being impaired (it did −50% in 2022 and −25% in a single day in Jan-2025). The realistic downside is a violent de-rating, not insolvency.

10. Valuation Discussion (Embedded Expectations)

The multiples are at a record on doubly-flattered earnings. At ~$712: trailing P/E ~38x (TTM GAAP EPS $18.85), ~45x on the company’s own adjusted EPS (~$15–16, stripping securities gains), EV/EBITDA ~28x, EV/sales ~3.8x, P/B ~13x. Every metric is the 98.6th–99.4th percentile of IES’s own decade (composite 98.9th) — its richest valuation ever. Critically, the GAAP P/E understates the richness: because GAAP EPS is inflated by ~$2–3/share of non-operating gains that are disappearing, the multiple on contracting earnings is ~45x — and on normalized mid-cycle earnings (if margins revert) it would be higher still.

Peer context. The closest analogs trade well below IES on the cleaner metrics: EMCOR ~20–26x earnings, Comfort Systems richer but on a more services-diversified base. IES — smaller, more housing-cyclical, less diversified, less recurring-revenue than EMCOR — trades at a premium to the higher-quality, more-diversified leader. That inversion (the more cyclical, more concentrated name at the higher multiple) is the valuation tell.

Embedded-expectations read — what ~$712 underwrites. At 38–45x peak-margin earnings, the market is pricing a multi-year continuation of peak conditions: hyperscaler capex sustained, the labor shortage holding 11%+ operating margins as a new structural floor, the data-center backlog converting without an air-pocket, and the M&A engine compounding — and it is implicitly forgiving the disappearing securities-gain tailwind and the housing drag. That is a demanding set of assumptions for a thin-moat cyclical contractor whose ROIC was 9% three years ago, layered on extreme momentum (beta 2.07) and a wholesale insider exit.

Scenario sketch (illustrative, no target):

• Bear (~$250–400): data-center capex digests or housing worsens; operating margin reverts toward mid-cycle (~6–8%); securities-gain prop gone; normalized EPS resets to ~$10–13; multiple compresses to a more cyclical ~20–25x. The stock has shown it can do −50%.
• Base (~$450–600): the super-cycle persists another 1–2 years at moderating growth; margins hold ~10–11% but stop expanding; adjusted EPS ~$16–19; the rich multiple compresses modestly toward ~30x as growth decelerates.
• Bull (~$800–1,000+): hyperscaler capex plateaus durably high and the labor constraint sustains 11%+ margins as a structural floor; EPS compounds toward ~$22–26 on backlog conversion + M&A; the market keeps paying a Constellation-style premium.

The risk/reward from ~$712 is skewed unfavorably: the bull requires peak conditions to become permanent, the base offers limited upside with multiple-compression drag, and the bear is a deep, demonstrated drawdown. No price target, no recommendation — these scenarios frame the embedded expectations only.

11. Variant Perception

Consensus belief: IES is a best-in-class, founder/operator-controlled data-center-infrastructure compounder — a “Constellation-for-the-trades” with a superb capital-allocation engine, record backlog, fortress balance sheet, and a multi-year AI/electrification runway. The +162% twelve-month run and record valuation reflect the market’s embrace of that narrative; the momentum cohort (EME/FIX/PWR/STRL) is all bid.

Strongest bull case: the data-center demand is genuine and durable, IES captures more of it per revenue dollar than diversified peers, the labor shortage structurally sustains elevated margins, the decentralized M&A model compounds capital at high returns, the balance sheet has dry powder, and a housing recovery is a free option. Own the jockey (Gendell/Tontine + Simmes) and let the super-cycle run; quality and demand compound through the multiple.

Strongest bear case: you are paying 38–45x peak-margin earnings near a record for a thin-moat cyclical contractor whose ROIC was 9% three years ago; ~20–25% of recent earnings was non-operating securities gains now disappearing as GIFI consolidates; ~39% of revenue is a housing-cyclical drag; the “moat” is a cyclical-peak-returns illusion plus a key-person jockey; comp has no returns governor and banks the investment gains; and — the loudest signal — the 53%-controlling owner and every top officer are selling ~$118M into the all-time high with zero buys. A data-center digestion or housing-led margin reversion resets normalized EPS toward ~$10–13 and compresses a 38x multiple violently; beta 2.07 means the drawdown would be brutal.

The 3–5 assumptions that matter most: (1) durability of hyperscaler/data-center capex (no digestion); (2) whether 11%+ operating margins are a structural floor or a cyclical peak; (3) the housing/Residential trajectory; (4) the post-GIFI earnings quality (how much normalized EPS remains ex-securities-gains); (5) whether the controlling owner’s distribution accelerates (resale-shelf overhang).

Factor-positioning evidence — the tape is the opposite of a falling knife; it’s the knife-thrower. IES carries an extreme momentum loading (+1.49), a strongly negative value loading (−0.96), beta 2.07, and rs_12m +162% — it sits within ~5% of its all-time high after a 28-bagger, in the top decile of momentum and the most expensive decile of its own history. This is a crowded, high-beta, late-stage momentum vehicle, not an abandoned value name. The variant-perception risk is therefore asymmetric: consensus, the tape, and the valuation are all aligned on continuation, so the under-priced scenario is any interruption — a capex pause, a margin wobble, a housing leg down, or simply the controlling owner continuing to sell. The market is right about the quality of the operator and the reality of the demand — but it is paying a permanent-peak price for a cyclical-peak moment, and the insiders are voting with their shares.

12. Fact vs. Interpretation Table

13. Open Questions

1. How durable is hyperscaler/data-center capex, and is IES’s record backlog pull-forward or run-rate? A digestion hits its two highest-margin segments simultaneously.
2. Are 11%+ operating margins a structural floor or a cyclical peak? The entire valuation rests on the answer; the FY22 2.6% margin is the cautionary precedent.
3. What is normalized, ex-securities-gains, mid-cycle EPS? Adjusted EPS is ~$15–16 at peak margins; a margin reversion could pull it materially lower.
4. Does the Tontine distribution accelerate? A 53%-owner on a resale shelf selling discretionarily at the high is a structural overhang of uncertain pace.
5. How accretive is GIFI once integrated (it began loss-making), and what is the run-rate margin of the consolidated Infrastructure Solutions segment without the marks?
6. Key-person risk — how much of the capital-allocation edge depends on Gendell (66, now Executive Chairman) vs the institutionalized model under Simmes?

14. What Must Be True

Bull case requires: (a) hyperscaler/data-center capex sustains multi-year, with IES’s backlog converting without an air-pocket; (b) the skilled-labor constraint durably holds operating margins at 11%+ as a new structural floor (not a cyclical peak); © the M&A engine keeps compounding capital at high returns; (d) a housing recovery turns the Residential drag into a tailwind. Falsification test: if, over the next 2–4 quarters, data-center backlog growth stalls or consolidated operating margin rolls back toward mid-cycle (~7–9%) while the securities-gain prop is gone, the “structural compounder at a fair-for-quality price” thesis breaks — normalized EPS resets toward ~$10–13 and a 38–45x multiple is unsustainable; the stock de-rates hard.

Bear case requires: (a) a data-center capex digestion or a housing-led downturn; (b) operating-margin mean reversion as labor eases or competition intensifies; © continued multiple compression from a record level; (d) ongoing insider distribution. Falsification test: if hyperscaler capex plateaus durably high, IES proves 11%+ margins are a structural floor across a full year, the post-GIFI operating earnings hold ~$16–19 adjusted EPS, and the insider selling stops (or an insider buys) — then the cyclical-peak bear is wrong, the Constellation-for-trades premium is validated, and the stock can grow into its multiple. The single cleanest signal either way is the trajectory of data-center backlog and consolidated operating margin over the next several quarters, read alongside the insider-transaction tape.

15. Source Appendix

See the Source Appendix below for the full primary-source list with URLs and dates. Principal sources: IES Holdings FY2025 Form 10-K (filed 2025-11-21); Q2-FY2026 Form 10-Q (filed 2026-05-01); DEF 14A (filed 2026-01-07); SC 13D/A (Tontine, 2026-01-09); the 8-K and Form 3/4 corpus (60-month mirror, including the May–June 2026 insider sales); ROIC.ai fundamentals/ratios/enterprise value; AZI price CSV and valuation-index percentiles; FactorsToday factor model.

The body of this article takes no investment position and contains no price target; the only position and valuation zone appear in the clearly-labeled Claude's Take block at the top, which is the author’s own subjective opinion and general information only — not investment advice.

APPENDIX A — Standard Diligence Questionnaire

IES Holdings, Inc. (NASDAQ: IESC) — Report date 2026-06-21

Supplemental diligence questionnaire. Answers labeled Fact / Interpretation / Assumption where it matters. Fiscal year ends September.

General

What thoughtful questions have other investors asked about this company? The recurring questions: (1) “Is this a Constellation-for-the-trades compounder or a lucky cyclical roll-up?” (answer: a genuinely good operator/allocator, but the 9%→34% ROIC arc is a cyclical-peak illusion, not a moat). (2) “How much of the earnings beat is real?” (~20–25% of recent pretax income is non-operating securities gains now disappearing as GIFI consolidates). (3) “Why are insiders selling so heavily?” (Gendell/Tontine + officers sold ~$118M into the ATH, discretionary, zero buys). (4) “Are 11% operating margins sustainable or peak?” (5) “What is the data-center concentration / digestion risk?”

Cyclicality & Earnings Nature

Cyclical high or low? Interpretation: unambiguously a cyclical (and momentum) HIGH — operating margin tripled from 2.6% (FY22) to 11.4% (FY25), ROIC went 9%→34%, the stock is up 162% over twelve months at its richest-ever valuation. The data-center complex is at a capex peak; the Residential segment is in a housing down-cycle (a partial offset).

Driven by external environment or internal action? Both, but predominantly external — the AI/data-center capex super-cycle plus a skilled-labor shortage that hands scaled contractors pricing power. Internal action (decentralized operating model, bidding discipline, M&A, prefab/labor productivity) is real and amplifies it, but the margin/ROIC explosion maps onto the external demand wave.

How stable are revenues? Project-based and cyclical — no recurring/subscription base (only a thin MSA/maintenance book). Backlog is at records (~$2.4B, +33% YoY) but ~29% is “agreements without an enforceable obligation,” and single-family/short-cycle work is excluded. Revenue can fall sharply in a downturn (FY22 proof).

Outlook / market size. Fact/Interpretation: the data-center/electrical-infrastructure market is large and growing (hyperscaler capex, reshoring, electrification), currently labor-constrained; the ~39% Residential business is housing-cyclical and currently contracting. International exposure minimal (US-centric).

Business Quality & Competitive Moat

Industry getting more or less competitive? The base MEP-contracting industry is permanently fragmented and competitive; the current moment is unusually favorable (demand > labor supply). High returns are attracting capital into the space (Marathon warning), with the labor bottleneck the only thing delaying mean reversion.

How profitable (ROIC/ROE)? Elite at peak: ROIC ~34%, ROE ~47% (FY25) — but ROIC was 9% in FY22. The high ROE is amplified by an asset-light, negative-working-capital model (ROE was ~47% even in FY20–21 when ROIC was ~17%).

How profitable is the industry / barriers to entry? Low barriers at the base (a two-truck shop is trivial to start); modestly higher at the complex/large-project tier (data centers, custom-engineered products). No proprietary technology, no network effects, minimal switching costs.

Easily understood? Yes — a decentralized holdco of electrical/mechanical contractors with four segments; the nuances are reading operating income (not GAAP EPS, which is inflated by securities gains) and recognizing the returns as cyclical-peak.

Undermined by foreign low-cost labor? No — field installation is local/onsite and not labor-arbitrage exposed; the binding constraint is the domestic skilled-trades shortage.

Do brands matter? Modestly — reputation/execution certainty matters in large data-center work (agency-relationship stickiness), but each project is competitively bid; there is no consumer brand moat.

Nature of competition / switching costs. Bid-to-bid competition; switching costs minimal. The durable edge, if any, is the capital-allocation model and the assembled skilled workforce — both real but a process/cyclical advantage, not a structural barrier.

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The skilled, trained workforce and the segment operating relationships are the real economic assets (unrecognized). Backlog (~$2.4B) is disclosed but partly non-binding.

Off-balance-sheet liabilities? Surety/performance bonds (standard for contractors), operating leases, fixed-price contract loss exposure, and the Tontine related-party office sublease. No financial-leverage risk (near net-cash). The Tontine resale shelf is a structural overhang.

How conservative is the accounting? Mostly clean (percentage-of-completion is standard; SBC tiny), with one important flag: GAAP EPS is inflated by volatile, non-operating “gain on marketable securities” (~$54M H1-FY26, ~20–25% of pretax income) — the company’s own adjusted EPS strips these out. Read adjusted/operating earnings.

How CapEx-hungry? Light — capex ~2% of sales (asset-light contracting). The recent GIFI deal adds a more capital-intensive (manufacturing) element to Infrastructure Solutions.

Capital Allocation & Management

FCF generation and use; philosophy. Strong (~$219M FY25 FCF), capital-light. Priority: M&A first, buybacks second, no dividend. Decentralized model with centralized capital allocation under Gendell.

Significant recent acquisitions? Gulf Island Fabrication (~$152M, Jan-2026, into Infrastructure Solutions — began loss-making); Broadwind Abilene facility (~$19.5M, May-2026); Greiner (~$67M, FY24) and tuck-ins. Disciplined on small deals; the largest recent deals are capacity/turnaround.

Buying back shares? Yes but mistimed — bought least when cheap (FY22–23 at ~$25–70), most when dear (FY24–25 at ~$170+); ~4.4% share reduction over 5 years; muted at the current peak.

Issuing large amounts of stock to insiders? No — SBC tiny (~$13M). The issue is the reverse: insiders are heavy sellers.

Compensation policy / motivations. Two red flags: (1) sole incentive metric is Adjusted Pretax Income — no ROIC/returns/per-share/TSR governor (rewards empire-building on a roll-up); (2) the metric does not exclude the marketable-securities gains, so non-operating investment marks pumped FY25 payouts to 129.2% of target. Insider behavior: Gendell/Tontine (~53%) + CEO/CFO/director sold ~$118M into the ATH (discretionary, non-10b5-1); zero open-market buys ever. Controlled company; say-on-pay effectively a formality.

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — a standard US C-corp common stock (NASDAQ: IESC); 1099, no K-1. Controlled company (~53% Tontine).

Dividend policy? None — never paid a dividend.

How profitable? Very, at peak — ~11.4% operating margin, ~34% ROIC, ~47% ROE (FY25); but single-digit ROIC three years ago.

Net income diverging from cash from operations? Modestly — FY25 FCF/NI ~0.94x (working-capital build in growth). The bigger divergence is within net income: ~20–25% of recent pretax income is non-cash/non-operating securities marks.

Risks & Downside

What would cause the stock to decline? A data-center capex digestion, operating-margin mean reversion, a deepening housing/Residential downturn, the vanishing securities-gain tailwind, multiple compression from a record level, or continued/accelerating insider selling. Beta 2.07 means any of these triggers a large move.

Risk of catastrophic loss? Low at the business level (near net-cash, asset-light, real demand) but high drawdown risk to the equity — a beta-2.07 momentum name at 38–45x peak earnings can lose 40–60% in a cyclical/multiple reset (it did −50% in 2022, −25% in one day in Jan-2025) without the business being impaired.

Chance of total loss? Remote — the balance sheet is near net-cash and the demand is real. The risk is a severe de-rating, not insolvency.

Recent News & Events

Business environment changed recently? Yes: (1) data-center demand drove a Q2-FY26 beat (+17.8% pop) and record backlog; (2) Residential softened (housing down-cycle); (3) GIFI consolidated (Jan-2026), removing the securities-gain tailwind; (4) leadership transition (Gendell → Exec Chairman, Simmes CEO, 2025).

Significant acquisitions? Gulf Island Fabrication (Jan-2026), Broadwind Abilene (May-2026). Accounting-policy changes? None material; note the securities-gains QoE flag. Other recent changes? Heavy insider selling (~$118M, May–June 2026); $200M buyback authorization (July-2024, lightly used).

APPENDIX B — Source Appendix

IES Holdings, Inc. (NASDAQ: IESC) — Report date 2026-06-21

Primary sources first. All facts trace to one of the public sources below. Price as of 2026-06-18 close. Fiscal year ends September.

Primary — SEC filings (trailing 60 months)

1. IES Holdings, Inc. Form 10-K, FY2025 — filed 2025-11-21 (period 2025-09-30). Decentralized-holdco model, four-segment detail (Communications/Residential/Infrastructure Solutions/Commercial & Industrial revenue, operating income, margins), backlog table, Tontine control, risk factors, revenue recognition. SEC EDGAR CIK 0001048268. https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001048268&type=10-K
2. Form 10-Q, Q2-FY2026 — filed 2026-05-01 (period 2026-03-31). Q2 revenue $974M (+17% YoY), dil EPS $5.44; the $37.3M “gain on marketable securities” (~25% of pretax); GIFI consolidation; Residential segment margin ~2%; RPO/backlog.
3. DEF 14A (proxy) — filed 2026-01-07. Tontine/Gendell beneficial ownership (~53.17%, the entity stack), comp design (sole metric Adjusted Pretax Income, no ROIC governor, 129.2%-of-target FY25 payout), CEO/Chairman pay, controlled-company governance.
4. SC 13D/A (Tontine) — filed 2026-01-09. Gendell/Tontine ownership detail.
5. Form 8-K corpus (33 filings, 2021–26) — quarterly earnings, GIFI acquisition (Jan-2026), Broadwind Abilene (May-2026), leadership transition (Gendell → Executive Chairman, Simmes CEO, 2025), buyback authorization (July-2024).
6. Form 3/4/5 corpus (221 Form 4s, 60 months) — the insider read: Gendell/Tontine + CEO Simmes + CFO McLauchlin + director Cleveland selling ~$118M (May–June 2026, $670–764, discretionary/non-10b5-1); zero code-P open-market buys.

Quantitative data sources

7. ROIC.ai — income statement, profitability ratios (ROE/ROIC/margins), cash flow, per-share data, enterprise value (EV ~$13.7B, EV/EBITDA ~28x at $712). Third-party aggregated; reconciled to filings.
8. AZI price CSV (azitrading.com/controls/download-data.php?t=IESC) — 5-year split/dividend-adjusted OHLCV; 5yr low $25.27 (May-2022), ATH $749.83 (Jun-12-2026), 52wk low $273.13 (Jun-2025), close $712.39.
9. AZI valuation-index — own-history percentiles: P/E 37.79x (98.6th), P/B 13.41x (98.7th), P/S 3.96x (99.4th), composite 98.9th = richest-ever.
10. AZI news feed — recent-events tape: Q2-FY26 data-center beat (+17.8%), heavy insider Form 4 selling (Gendell/officers), Broadwind/GIFI deals.
11. FactorsToday factor model — beta 2.07, alpha +0.95; rs_12m +162.45%, rs_peak −4.99%; loadings (Momentum +1.487, Market +1.412, SmallSize +0.999, Value −0.959, Industrials +1.066, Infrastructure +0.797); leaderboard (m3 ann +606%/Sharpe 9.1, y1 +159%, y10 +49% ann, maxDD y5 −54%); factor-twins FIX/EME/STRL/PRIM/PWR/ROAD/NVT.

Secondary — industry / market data

12. Hyperscaler capex / data-center demand data (public cloud-platform capex disclosures, industry estimates) — the demand driver for Communications + Infrastructure Solutions.
13. US housing-starts / NAHB data — the Residential-segment cycle context.
14. Broadwind / GlobeNewswire (2026-05-05) — Abilene facility asset-acquisition terms.
15. Trade press / company IR — IES holds no live earnings calls (Tontine low-IR model); forward framing sourced from prepared CEO/Chairman quotes in quarterly press releases plus 10-K/10-Q MD&A.

Note: ROIC.ai, AZI, and FactorsToday are third-party aggregated/statistical sources, not primary; for US-filer facts, EDGAR and the 10-K/10-Q are authoritative and were used to reconcile every material number.