GE Vernova

Phase A — Understand the business

Company Overview

GE Vernova is the energy business spun out of General Electric on April 2, 2024 (GE's remaining aviation unit became GE Aerospace). It is a pure-play electric-power franchise organized in three reportable segments:

• Power — FY2025 revenue $19,767M. Gas Power ($16,006M) is the crown jewel: heavy-duty (HA-class) and aeroderivative gas turbines plus a large, sticky services book. Also Nuclear ($1,018M), Hydro ($806M) and Steam ($1,937M).
• Wind — FY2025 revenue $9,110M. Onshore Wind ($8,241M), Offshore Wind ($652M, shrinking and loss-making) and LM Wind Power blades ($217M).
• Electrification — FY2025 revenue $9,642M. Grid Solutions ($6,620M — transformers, HVDC, switchgear, AC substations), Power Conversion & Storage ($2,049M), Electrification Software ($973M). The fastest grower (+28% YoY).

Business model. A capital-goods + long-life-services franchise. GEV sells big iron (turbines, grid gear) on multi-year project contracts, then earns decades of high-margin parts and service revenue on the ~7,000-unit installed gas-turbine base (~1,800 units under long-term service agreements, ~10-year average remaining contract life). The installed base is the moat and the annuity. Roughly 45% of FY2025 revenue is services ($17,134M of $38,068M).

Contract structure & payment terms. Equipment is increasingly sold with slot reservation agreements (SRAs) — customers pay upfront to lock future gas-turbine production slots: 29 GW reserved at YE2024 → 43 GW at YE2025. These advance collections are the reason operating cash flow ran far ahead of earnings (an $8,019M FY2025 working-capital inflow from "progress collections and current deferred income") — a structurally favourable, customer-funded growth dynamic.

Customers. Utilities, independent power producers, grid operators, and — the new marginal buyer — hyperscalers and data centers driving AI electricity demand. Scale: ~75,000 employees (~70% in manufacturing/engineering/services); HQ Cambridge, MA; CEO Scott Strazik.

Supply Chain

Upstream inputs → GE Vernova → end customer, named at each node:

• Upstream raw inputs: specialty steel and forgings, nickel-based superalloys and single-crystal castings (hot-section turbine blades), copper and grain-oriented electrical steel (transformers/grid), rare-earth magnets and balsa/glass/carbon for wind blades. Grain-oriented electrical steel and large-transformer capacity are the binding industry chokepoints.
• Tariff exposure: global tariffs cost GEV ~$250M in FY2025 after contractual protections and mitigation — a manageable but rising input-cost line.
• The company (transformation): heavy-duty turbine manufacture (Greenville SC for HA-class), aeroderivatives, LM Wind Power blade plants (incl. Gaspé, Quebec — see Lens 10), and — newly in-house — Prolec GE transformers (acquired Oct 2025, closed Feb 2026; ~10,000 employees, seven plants, five in the US). Vertically integrating transformers is a direct response to the grid-equipment bottleneck.
• Downstream / end customers: utilities (NextEra, Southern, Duke-type IOUs), IPPs, national grid operators, and hyperscalers (Microsoft, Amazon, Google, Meta, xAI-type loads) buying behind-the-meter and grid-tied gas + grid gear for AI campuses.

Chokepoints / single-source dependencies: (1) GEV's own HA-turbine production capacity is the bottleneck — demand exceeds slots, which is why SRAs exist; (2) industry-wide large-power-transformer and grid-steel scarcity (the reason for the Prolec deal); (3) the global gas-turbine hot-section supply chain (superalloy castings) is concentrated and long-lead. This lens passes the "names or it didn't happen" bar: the binding constraints are GEV's Greenville turbine line and the transformer supply base it is buying into.

Competitive Advantages (moats)

Named competitive set (from the 10-KForm 10-KA company’s audited annual report to the US regulator. The most complete thing it publishes.):

• Power: Siemens Energy, Mitsubishi Power, Westinghouse, Framatome, Rolls-Royce
• Wind: Vestas, Siemens Gamesa, Nordex, Envision, Goldwind
• Electrification: Hitachi Energy, Siemens Energy, Siemens, Schneider Electric, Mitsubishi Electric, ABB

The moat is real and concentrated in gas + services. Heavy-duty gas turbines are a true three-firm global oligopoly — GE Vernova, Siemens Energy, Mitsubishi Power — with backlogs stretching toward ~2030. Durable moat sources:

1. Installed-base lock-in / switching costs — ~7,000 turbines, ~1,800 under LTSAs, ~10-yr average life. Once a utility runs GEV iron, it buys GEV parts and service for decades. This is the highest-quality earnings in the company.
2. Capacity scarcity = pricing power — with demand exceeding turbine slots, GEV captured favourable price across Gas Power in FY2025 (an explicit driver of the +18% organic Power EBITDA), and SRAs let it bank pricing years forward. The bargaining power has inverted toward the supplier — buyers now pre-pay to reserve.
3. Technology / IP — HA-class turbine efficiency leadership; HVDC and large-transformer know-how in Grid Solutions; nuclear (BWRX-300 SMR optionality via GE Hitachi).
4. Scale + global service footprint — 75,000 employees, global plants, hard to replicate.

Bargaining power: strong over customers in gas/grid today (scarcity); weaker in onshore wind (commoditized, policy-dependent); the company is the price-maker in the parts of the portfolio that matter.

Segments

All figures FY2025 / FY2024 / FY2023, USD millions:

Segment-of-segment trends: Gas Power $16,006M (+11%) is the engine; Electrification's Grid Solutions ($6,620M, +34%) and Power Conversion & Storage are the second leg. Wind is bifurcated — Onshore ($8,241M) is finally improving on price; Offshore Wind ($652M, down from $1,377M) and LM Wind Power ($217M, down from $542M) are the value-destroyers, dragged by a US offshore-lease pause and footprint reduction.

Geography: workforce split is the best public proxy — ~24,000 Europe, ~21,000 US, ~19,000 Asia, ~6,000 Latin America; the 10-K does not disclose a clean three-segment geographic revenue split, so a precise geo-revenue table is n/a at segment granularity.

Backlog (RPO) — the leading indicator:

• Total RPO $150.2B (YE2025) vs $119.0B (YE2024) = +26%. Equipment $64.2B / Services $86.0B.
• By segment: Power $94.4B (+29%), Electrification $34.7B (+48%), Wind $21.6B (−5%).
• Then $163.3B at Q1-2026 (vs $123.4B a year prior) — still accelerating. Management targets ~$200B by YE2028.

---

Phase B — Measure performance

Earnings Result (latest print: Q1-2026, reported 2026-04-22)

Headline GAAP numbers:

• Revenue $9,339M, +16% YoY (Q1-25: $8,032M). Equipment $5,254M / Services $4,084M.
• GAAP net income $4,742M, net margin 50.9%, diluted EPS $17.44.
• Adjusted EBITDA $0.9B (+$0.4B YoY); operating income only ~$0.1B.
• Free cash flow $4.8B (vs $1.0B) — again advance-collection-driven.
• RPO $163.3B; orders +71% YoY.

Read the GAAP number with extreme care. The $4.7B net income is overwhelmingly non-operating: a $3,992M pre-tax remeasurement gain from revaluing GEV's previously-held 50% Prolec stake to fair value on consolidation, plus a ~$0.3B Proficy (software) disposal gain — together ~$4.3B of one-time items, vs underlying operating income of ~$0.1B. The honest operating number is the adjusted EPS of $2.06, beating ~$1.88 consensus.

• Margin moves: Adjusted EBITDA roughly doubled on Power/Electrification price + volume; Prolec consolidates into Electrification ("Power Transmission" now appears as a unit).
• Guidance: post-Q1, GEV raised FY2026 to ~$44.5-45.5B revenue (from $41-42B), with 11-13% adjusted EBITDA margin and $4.5-5.0B FCF.
• Balance-sheet flags: cash $8.8B, essentially net-cash pre-Prolec; $3.0B revolver; the Prolec close uses ~$5.3B (half cash/half debt) so net leverage steps up modestly from a near-zero base.
• Market reaction: stock surged ~14% on the Q1 print — the market rewarded orders/backlog and the guide-raise, not the optical $17.44 EPS.

FY2025 context: revenue $38,068M (+9%); GAAP net income $4,879M (margin 12.8%), diluted EPS $17.69 — but flattered by a ~$2.9B US tax valuation-allowance release (effective tax rate −72.5%; a $2,051M tax benefit). Adjusted EBITDA $3,196M (8.4% margin) is the clean figure; gross margin 19.8% (up from 17.4%); FCF $3,710M.

Earnings Calls (sentiment trend)

No transcripts on the local shelf (transcripts/ empty); sentiment is drawn from filings MD&A tone + investor-update coverage, labelled /.

Trajectory of management tone (FY2024 → Q1-2026): steadily, then aggressively, more confident.

• The recurring, escalating themes: "electricity demand from hyperscalers and data centers," gas as "reliable and dispatchable" baseload, pricing discipline, services-margin expansion, and "lean operating model" productivity.
• The clearest sentiment signal is the Dec 2025 investor update: management raised the 2028 framework materially, doubled the dividend, and lifted the buyback to $10B — a capital-return posture only a team confident in forward cash flow takes.
• What they stopped emphasizing: the energy-transition/decarbonization framing has visibly receded relative to the reliability/AI-demand/dispatchable-baseload framing — the narrative pivoted from "green transition" to "powering AI." Offshore wind has gone from growth story to a managed wind-down ("fewer and more reliable workhorse products," footprint reduction).

Comps

GEV + key listed peers. Multiples are with source/date or n/a.

Takeaway: on the data that is sourceable, GEV trades at a clear premium — forward P/E ~52-60x vs Siemens Energy ~31-35x EV/EBITDA — justified by faster growth, higher-quality gas/services mix, a US-centric footprint, and a cleaner balance sheet, but it is unambiguously a priced-for-perfection multiple. Where a peer multiple could not be cleanly sourced at run time it is left n/a rather than fabricated.

Stock-Price Catalysts (moves >5%, what the market actually reacts to)

GEV has only traded since April 2024, so the "5-year" window is its full 2-year life. The tape is up enormously — 52-wk range $482.20-$1,181.95, ATH $1,181.95 on 2026-04-23, price **$1,110 (2026-06-29)**, +117% over the trailing year (and up ~240%+ from spin lows on longer windows).

What moves it (pattern):

1. Order/backlog prints and guidance raises — the +14% pop on Q1-2026 (orders +71%, guide-raise) is the template. The market reacts to forward demand visibility, not trailing EPS.
2. The Dec investor updates — the multi-year framework raises (2025: $52B/20%/$22B) are major re-rate events.
3. AI-power demand headlines — anything tightening the data-center → gas-turbine → grid narrative (hyperscaler power deals, capacity shortages) lifts the name.
4. Offshore-wind / policy negatives — the Dec 22, 2025 US Interior offshore-lease pause and the Vineyard Wind dispute are the recurring downside triggers, though contained to a small, already-impaired segment.

The signal: this is a momentum/expectations stock keyed to orders and the AI-power thesis, not to quarterly accounting EPS.

---

Phase C — Judge people & books

Management

• CEO Scott Strazik — 20+ years at GE; ran GE Gas Power from 2018, expanded to lead GE Power in 2021, then CEO of standalone GE Vernova from the 2024 spin. He is the architect of the gas-turbine turnaround that is now the company's profit engine — a strong, operationally credible track record in the exact business that matters.
• Track record (quantified): Power segment EBITDA margin went 9.9% (2023) → 14.7% (2025); Electrification 3.7% → 14.9%; total Adjusted EBITDA $807M → $3,196M in two years; gross margin 14.5% → 19.8%. The operating improvement is real and broad.
• Capital allocation — disciplined and shareholder-friendly: doubled the dividend (to ~$0.50/qtr / $2.00 annualized) and raised the buyback authorization to $10B (from $6B), having repurchased 8.2M shares for $3.3B in FY2025. The Prolec GE buy-in ($5.3B) is a strategically logical vertical-integration of the scarce transformer supply chain — though paying up at a cycle peak is a fair critique. Net: reinvest + return, from a near-net-cash base.
• Skin in the game: founder-equivalent insider ownership is modest (this is a spun-out division, not a founder company); our figures is not on the shelf, so precise insider holdings are n/a.
• Red flags: the LM Wind Power (Gaspé) data-falsification allegation — an internal investigation found employees were instructed by supervisors to falsify QA data on wind blades — is a genuine governance/quality-culture flag, though confined to the offshore/blade unit. No related-party self-dealing or comp scandals surfaced.
• Archetype: professional operator/turnaround manager, not founder-visionary — the right archetype for a scale industrial harvesting a cyclical upturn.

Forensic Red Flags

Forensic-analyst read across the three statements — every figure labelled.

1. GAAP earnings are heavily flattered by non-operating items — the #1 flag. FY2025 net income $4,879M includes a ~$2.9B tax valuation-allowance release (ETR −72.5%, a $2,051M tax benefit). Q1-2026 net income $4,742M includes a $3,992M Prolec remeasurement gain + ~$0.3B Proficy gain. ~$7B of the trailing-five-quarter GAAP net income is one-time. Anyone valuing GEV on reported P/E ($17.69 + $17.44 ≈ $35 TTM "EPS") is anchoring on a number that does not recur. Use Adjusted EBITDA (8% margin) and adjusted EPS ($2.06 Q1) instead.
2. Operating cash flow >> earnings, but for a fragile reason. FY2025 CFO $4,987M vs operating income $1,388M, driven by an $8,019M working-capital inflow from progress collections / deferred income. This is good (customer-funded, SRA-driven) but cyclical — it reverses if orders decelerate. FCF quality is order-momentum-dependent, not steady-state.
3. Wind contract losses & remeasurement. Offshore Wind carries chronic losses; the segment has booked contract losses and supply-agreement terminations. Watch for further charges as the US offshore book is worked down.
4. Acquisition accounting to scrutinize going forward. The Prolec purchase added $3,980M of intangibles on a $9,656M total consideration; future intangible amortization and any goodwill impairment risk now sit in Electrification.
5. SBC and non-GAAP gap — SG&A rose partly on higher stock-based comp; the company leans on Adjusted EBITDA, but the adjustments here are mostly removing favourable one-timers (good discipline) rather than hiding costs.

Regulatory findings (required):

• SEC: No SEC Litigation Releases or AAERs name GE Vernova (verified via EDGAR EFTS, LR + AAER, 2021-06-30 to 2026-06-30).
• Non-SEC enforcement: web search surfaced no DOJ/FTC/SEC enforcement action. The material legal exposure is civil/contractual: Vineyard Wind sued GEV (April 2026) to block it from exiting the supply agreement, withholding payment to offset $853M in costs from defective blades; a Massachusetts judge granted Vineyard Wind a preliminary injunction/TRO (April 2026) preventing termination. Separately, blade failures at Dogger Bank (UK) and Vineyard Wind (US) were attributed to a "manufacturing deviation"/insufficient bonding, and a data-falsification allegation at the Gaspé blade plant is under internal investigation. Dec 22, 2025: US Dept of Interior paused leases for all large-scale offshore wind under construction, hitting Vineyard Wind's timeline.
• 10-K Item 3 (Legal Proceedings): the FY2025 10-K's Item 3 is a brief cross-reference (legal-proceedings detail is carried in the financial-statement notes/contingencies rather than a long Item 3 narrative); no single litigation is flagged as individually material to the consolidated financials beyond ordinary-course and the wind disputes above.
• Net: No accounting-fraud or securities-enforcement history. The real legal/quality risk is isolated to Offshore Wind/blades — a small, loss-making, shrinking part of the company — not the gas/grid core.

---

Phase D — Project & stress-test

Forward Projection (FY2026 / FY2027 / FY2028)

Built bottom-up from management's own raised framework + FY2025 actuals; every input labelled, output.

Anchors:

• Mgmt FY2026 guide (post-Q1 raise): revenue ~$44.5-45.5B, adj EBITDA margin 11-13%, FCF $4.5-5.0B.
• Mgmt 2028 framework: revenue ~$52B, adj EBITDA margin ~20%, cumulative FCF ≥$22B (2025-28).
• FY2025 base: revenue $38.07B, adj EBITDA $3.20B (8.4%), ~272M diluted shares.

Adjusted-EPS path (the operating number, ex-one-timers):

• Q1-2026 adjusted EPS $2.06. Annualizing the operating ramp and applying management's margin trajectory:
• FY2026 adj EPS ~$9-11. Brackets the messy $14.33-$30.93 GAAP-contaminated street range — the clean operating line is materially below the GAAP-inflated estimates.
• FY2027 adj EPS ~$13-16.
• FY2028 adj EPS ~$18-22.

Scenarios:

• Base: mgmt hits ~$52B / ~20% by 2028 → FY2028 adj EPS ~$20; at a (de-rated) ~30x → ~$600 fair value — below today's ~$1,110, i.e. the stock already discounts the 2028 plan and more.
• Bull: AI-power demand pushes revenue past $55B with >20% margin and sustained SRA pricing → adj EPS ~$24+; multiple holds ~40x → >$950.
• Bear: orders decelerate, working-capital tailwind reverses, a gas-cycle pause + offshore charges cap margins ~14-15% → adj EPS ~$12-14; multiple compresses to ~20-25x → ~$280-350.

(Brier forecast create step deliberately skipped — this is the --watchlist unattended loop; per SKILL.md a forecast is logged only on a genuinely committed base case in an interactive pass.)

Bull vs Bear

Bull case. GE Vernova is the best-positioned Western pure-play on the most durable macro tailwind of the decade — electrification + AI-driven power demand. It sits in a true gas-turbine oligopoly with capacity scarcity it can price against; a $150-163B backlog (and climbing to a $200B target) gives multi-year revenue visibility; SRAs let it bank pricing years forward with customer-funded cash; the installed-base services annuity is high-margin and sticky; Electrification rides the grid-capex supercycle; and a near-net-cash balance sheet funds a doubled dividend, a $10B buyback, and the value-accretive Prolec integration. Margins are inflecting (Adj EBITDA 8.4% → 20% targeted by 2028) and management has out-delivered and out-guided every quarter since the spin.

Bear case (permanent-impairment risks):

1. Valuation, not business quality, is the risk. At 52-60x forward earnings the market has priced flawless execution of the 2028 plan; my base-case fair value ($600 at a de-rated multiple on ~$20 FY2028 adj EPS) sits below the current ~$1,110. The asymmetry is poor at this entry.
2. Cyclicality dressed as secularity. Gas-turbine demand is historically boom-bust; the current order/SRA surge and the $8B working-capital tailwind reverse hard if the AI-capex cycle cools, datacenter power plans get rationalized, or rates/financing tighten utility capexCapital expenditureMoney spent on long-lived things — buildings, machines, servers — rather than on running costs..
3. Margin durability is unproven at 20%. The 2028 target assumes pricing power and productivity hold through inflation, tariffs (~$250M and rising), and labor. Power is a long-cycle, fixed-price-contract business where a few mispriced large projects can erase quarters of gains (Wind is the live cautionary tale).

Pre-mortem (18 months out, thesis broke): AI-datacenter power demand normalized as efficiency gains and grid-interconnect bottlenecks slowed buildouts; a couple of hyperscalers paused/renegotiated gas commitments; orders decelerated and the advance-collection cash engine reversed; an offshore-wind charge plus the Vineyard Wind judgment hit; the multiple compressed from ~55x toward ~25x and the stock halved — with earnings still growing. The break was the multiple, not the franchise.

Contrarian view (what the market is refusing to see): the bulls treat the SRA backlog as locked recurring revenue, but it is cyclical advance-pay equipment demand — the highest-quality earnings (services annuity) are a minority of profit, while the part driving the re-rate (equipment pricing) is the part that historically mean-reverts. The market is paying a software-like multiple for a capital-goods cycle.

Devil's Advocate (short-seller)

Dismantling the bull case.

• What structurally breaks it: the entire re-rate rests on AI-power demand staying vertical. If datacenter electricity-demand forecasts prove overbuilt (the same way every prior "demand supercycle" — nuclear, coal, shale-power — overshot), GEV's order book and SRA pipeline deflate, and the $8B/yr working-capital tailwind becomes a headwind.
• Concentration: revenue is concentrated in Gas Power ($16B, the profit center) and increasingly in a single demand vector (hyperscaler/AI power). A pullback by even two or three mega-buyers reorders the narrative.
• Why the moat is weaker than bulls think: Siemens Energy is a credible, well-backlogged co-oligopolist now executing its own turnaround; in a demand cooldown the three players compete on price again, and gas-turbine pricing power evaporates faster than bulls model.
• Worst capital-allocation read: buying the rest of Prolec at a cycle peak (~$5.3B, $3.98B of intangibles) and accelerating a $10B buyback at ~55x earnings — repurchasing your own stock near an all-time high is value-destructive if the multiple normalizes.
• Accounting that flatters: ~$7B of one-time gains (tax-allowance release + Prolec remeasurement) make GAAP EPS look ~3-4x the real operating number; a casual screen on reported P/E badly understates how expensive this is.
• Price test: if revenue/margins disappoint by 20-30% vs the 2028 plan (say $42B at ~15% adj EBITDA), adj EPS lands ~$12-14 and the stock is worth ~$300-350 at a normalized 25x — a ~70% drawdown from ~$1,110 with the business still growing.
• The single permanent-impairment scenario: a multi-quarter AI-datacenter capex pause coinciding with a large mispriced gas-project loss and a final offshore-wind blowup — plausible, perhaps 15-20% over 2-3 years, and it would re-rate the stock to mid-cycle-industrial multiples for years.

Management Questions (ordered by information value)

1. Of your $163B RPO, how much is firm equipment vs cancellable, and what share of FY2028's $52B revenue target is already in backlog vs still-to-be-won?
2. What portion of the 43 GW of SRAs carries non-refundable deposits and firm pricing, and what is the cancellation/penalty structure if a hyperscaler walks?
3. Walk us through the bridge from 8.4% Adjusted EBITDA margin (2025) to 20% (2028) — how much is price, mix (services vs equipment), volume leverage, and productivity, and which is most at risk?
4. How concentrated is gas-turbine demand among hyperscalers/datacenters, and what is your largest-customer and top-5-customer revenue concentration today?
5. What gas-turbine slot capacity (GW/yr) can Greenville actually deliver in 2027-2028, and what capex is required to hit it without eroding returns?
6. How should we think about the durability of current gas-turbine pricing through a demand pause — what happened to pricing in the last gas downcycle, and what's structurally different now?
7. What is the realistic terminal state and total remaining cash cost (charges, the Vineyard Wind $853M dispute, lease pause) of winding down Offshore Wind?
8. What did the Gaspé QA data-falsification investigation conclude, what remediation is in place, and what is the residual warranty/liability exposure across the blade fleet?
9. On Prolec at ~$5.3B with ~$4.0B of intangibles at a cycle peak — what return hurdle did it clear, and what synergy/margin accretion underwrites it?
10. With the stock near all-time highs at ~55x earnings, why is a $10B buyback the best use of capital versus reinvestment, M&A, or balance-sheet flexibility?
11. How exposed is the 2026-2028 plan to tariffs and input-cost inflation (grid steel, superalloys), and how much is contractually passed through vs absorbed?
12. What is the normalized, mid-cycle Adjusted EBITDA margin for each segment once the current order surge moderates?
13. How much of FY2025-26 free cash flow is the advance-collection/working-capital tailwind, and what does steady-state FCF conversion look like once orders normalize?
14. What is the realistic revenue and margin contribution timeline for nuclear/SMR (BWRX-300), and how much capital are you committing before it is cash-generative?
15. If AI-datacenter power-demand forecasts are revised down 25%, which parts of the plan are protected by the installed-base services annuity and which are fully exposed?

---