Layer 7 · GEV — GE Vernova
One-line thesis
GE Vernova is the power-infrastructure pure-play with $163 billion backlog, +71% organic order growth and FY26 guide raised — own it as the AI-power-bottleneck call that consensus is still under-modelling versus the gas-turbine slot reservation pipeline.
What GE Vernova physically does
GE Vernova is the spun-out power-equipment arm of the old GE — gas turbines, steam turbines, wind turbines, grid-electrification equipment (transformers, switchgear, HVDC converters), and the digital/services layer that supports the installed base. Inside the AI-stack story, the relevant businesses are the gas turbine business (which is supplying the new combined-cycle gas plants being built specifically for hyperscaler data centre power) and the Electrification segment (which is supplying transformers, switchgear and grid equipment for the data centre interconnection build-out).
The AI-stack layer is Layer 7 — power infrastructure — and it is the deepest binding constraint in the entire AI capex stack right now. You can have all the GPUs, HBM, transceivers and substrates you want; if you cannot get a 100 MW grid connection and the gas turbine slot to back it up, the data centre does not run. The US grid interconnection queue is now backed up to 2029-2031 in most ISO regions. Gas turbine lead times have stretched from 18 months in 2022 to 4-5 years in 2026. GEV is the principal global supplier of utility-scale gas turbines and is one of the few names where the supply-side constraint is literally physical capacity rather than financial willingness.
The Q1 26 disclosure that Electrification booked $2.4 billion in data centre orders in a single quarter — more than the full year 2025 captures the inflection. The data centre power demand curve has gone vertical, and GEV's order book is the financial signature of that curve.
The financial print
Q1 2026 (reported April 22) printed revenue $9.34 billion, EPS consensus beat, orders +71% organic YoY, backlog $163 billion (+$13 billion QoQ), and management raised FY26 revenue guide to $44.5-45.5 billion and FCF guide to $6.5-7.5 billion. The Electrification segment's $2.4 billion data-centre order print in Q1 alone was the headline number that re-rated sell-side models.
The backlog math is structural. $163 billion versus $116 billion at the spin (less than two years ago) is a 41% backlog expansion in 18 months, on a business where backlog converts to revenue over 2-5 years depending on segment. Gas turbine slot reservations are now being booked through 2029-2030 and management has guided to 110 GW of slot reservations targeted by year-end 2026. The stock closed $1,049.23 on May 17, up roughly +145% one-year total return. Q2 2026 reports July 22.
The orders +71% organic YoY metric deserves separate attention. Organic order growth at that rate, on a $40+ billion revenue base, is unprecedented in modern utility-equipment history. The closest analogue is probably the early-2000s combined-cycle gas-turbine cycle that lifted GE Power orders by 50-60% YoY for two consecutive years. That cycle ran for roughly five years before saturating; the current AI-driven order cycle has plausible duration through 2030-32 based on grid-interconnection queue depth and gas-turbine slot-reservation forward bookings. The historical comparison suggests the order book has another two to three years of structural expansion ahead before mean reversion sets in.
Customer mix
Customer mix is unique among the Tier A names — it is principally utility companies (Duke, NextEra, Southern, Entergy, Dominion, plus the Canadian, European and Middle Eastern utility customers), independent power producers, and the hyperscalers' own captive-power vehicles. The customer book is also unique in that the contracting is long-dated (5-15 year purchase agreements with deposit structures) and the cancellation costs are prohibitive, which makes the backlog quality very high.
The Middle Eastern utility customer base is increasingly material to the GEV story. Saudi Arabia's National Center for AI Capability (NCAC) is committing tens of billions of dollars to data centre build-out across the kingdom, paired with new gas-turbine generation to support the load. UAE's G42 and the broader Mubadala AI infrastructure programmes are doing the same at slightly smaller scale. GEV is a principal supplier to both regions through long-standing utility relationships with SEC (Saudi Electricity) and EWEC (UAE). The Middle Eastern AI-power capex is an under-appreciated incremental driver because most US-focused sell-side models do not properly capture the regional contribution to GEV's order book.
The hyperscaler-driven demand is funneled through three channels: direct captive-power purchases (Amazon's nuclear+gas+renewables ladder, Microsoft's Three Mile Island restart deal, Google's geothermal+SMR programmes), utility-led data-centre interconnection (where the utility procures GEV equipment on behalf of the hyperscaler load), and IPP-driven (where merchant generators like Vistra, Constellation, Calpine procure GEV equipment to serve hyperscaler PPAs). All three channels are accelerating, and the second-and-third channels are arguably even more durable than the first because the utility and IPP customer base is regulated and capital-disciplined in a way that hyperscaler capex is not.
The customer-quality angle is unusual for an AI-themed name. Unlike GPU/HBM/transceiver customers — which are concentrated in a handful of hyperscalers and a smaller tier of AI-cloud merchants — GEV's customer base is broadly diversified across hundreds of utility, IPP and industrial accounts globally. The hyperscaler-driven incremental demand is layered on top of a structural utility-replacement base that does not depend on AI capex continuing. Even in a hypothetical scenario where AI data centre demand collapses entirely, the gas-turbine fleet replacement cycle alone supports 15-18 GW/year of new build globally; the AI-driven incremental is what takes the curve to 25-30+ GW/year. The downside cushion is structurally larger than for any other Tier A name.
Competitive context
The competitive set in utility-scale gas turbines is a global oligopoly: GEV, Siemens Energy, Mitsubishi Power, and to a lesser extent Ansaldo Energia. The four together supply >95% of new utility-scale gas-turbine capacity globally. GEV has roughly 35-40% global market share and is the technology leader in the HA-class (high-efficiency advanced) gas-turbine segment that dominates new combined-cycle builds. Siemens Energy is the credible #2 and has been ramping its order book concurrently.
The competitive moat is the installed base (GEV's installed turbines globally generate ~30% of world electricity, providing decades of service revenue), the technology depth on combustion efficiency, and the manufacturing scale at the Greenville, SC and Belfort, France factories that competitors cannot easily match. The competitive context is structurally favourable through the AI-power-bottleneck cycle and probably extends through the next decade given the grid build-out requirements.
The Electrification segment is the second-engine story that gets less attention than the gas-turbine line but is structurally just as important. Transformers, switchgear, HVDC converters and grid-interconnection equipment are the second tier of the AI-power-bottleneck — even after you have the generation capacity, you need the equipment to get the power from the substation into the data centre. The market for utility-grade transformers and switchgear has tightened to the point where lead times for major substation equipment are now 36-48 months, up from 12-18 months pre-pandemic. GEV is the principal global supplier of this equipment alongside Siemens, ABB, Schneider Electric and Hitachi Energy. The $2.4 billion of data-centre orders booked in Q1 2026 alone — more than the entire 2025 — captures the Electrification inflection.
Terminal risk
The terminal risk is the AI capex blink translated into the slowest-moving part of the supply chain. If hyperscaler capex blinks in 2027-28, the immediate effect is GPU and memory orders rephased — the gas turbine orders, with multi-year lead times and committed deposits, are stickier. But over a 3-5 year window, sustained AI capex moderation would cool the gas-turbine order book and the +71% organic growth would mean-revert to the 8-12% structural utility CAGR.
The secondary terminal risk is the execution risk on slot oversubscription. 110 GW of slot reservations by year-end 2026 is a massive scaling-up for a manufacturing footprint that delivered 15-20 GW/year historically. Quality issues, capex bottlenecks, or supply-chain failures during the ramp would erode margin and damage customer relationships. Management has flagged this risk; it is real but historically GE Power has navigated similar ramps reasonably well.
A third terminal risk is regulatory and policy disruption. Gas-turbine combined-cycle plants are increasingly contested at the state and federal regulatory level — particularly in California, the Pacific Northwest, and the EU — where emissions rules and decarbonisation mandates could constrain new-build approvals. If regulatory friction increases meaningfully in 2027-28, the gas-turbine order book could face slot cancellations even if AI demand is intact. The mitigation is that the bulk of GEV's gas-turbine order book is in the US South-East, Texas, and the Middle East — geographies with materially more permissive regulatory profiles.
Bull case
The three-to-five-year bull case sees GEV converting the $163 billion backlog into $50+ billion of annual revenue by FY28, gas-turbine deliveries running 25-30 GW/year by 2028, Electrification revenue running $18-20 billion (versus FY26 closer to $13 billion), and operating margin expanding to the mid-teens consolidated. FCF in the $10-12 billion range. EPS in the $30-35 range. A 30x multiple — power-infrastructure monopolist with secular AI bottleneck moat — gets the stock to $900-1,050 on a 24-month view and $1,300-1,500 on a 36-month view.
The stretch bull case adds the small-modular-reactor (SMR) joint venture programmes maturing — GEV has positioned through BWXT and other strategic partnerships on the next-generation nuclear technology — plus the European grid build-out (REPowerEU, German Energiewende) layering on incremental orders independent of US AI demand. That path takes the stock to $1,800+ over five years.
The services revenue compounding is the often-overlooked third leg of the bull case. GEV's installed-base service business generates 35-40% operating margins on recurring multi-year contracts and grows at 6-8% CAGR independent of new-build orders. As the new-build cycle peaks and rolls off, the services line continues to compound, providing earnings durability and a high-quality-of-earnings anchor that pure capital-equipment names lack. By 2030, services revenue could be running $12-15 billion against $4-5 billion of operating income from services alone — a meaningful slice of consolidated earnings that does not depend on AI capex.
Gap / bear case
What the market may be missing is the durability of the backlog conversion. Gas-turbine deliveries through 2030 are essentially price-locked through the existing slot reservations and the deposit structures make cancellation prohibitive. Consensus FY27-28 revenue at $48-50 billion probably understates the conversion velocity if the manufacturing ramp executes.
The bear case the market may be under-pricing is the +145% one-year valuation expansion. The stock has moved from a re-rating story to a fully-priced compound story, and at ~30x EBITDA on consensus 2026 the entry-point asymmetry is less favourable than it was 12 months ago. Price has softened from $1,275 to $1,049 over the past eight sessions on broader power-sector profit-taking — the operating story is durable but the technical positioning is heavy, and unlike CEG and VST there is no STRONG_EXIT signal flagged in the May 11 daily brief for GEV, so this remains a buy-the-pullback setup rather than a tactical-downgrade setup.
A specific bear-case scenario worth modelling is a wind-turbine segment write-down. GEV's wind business has been a periodic drag on consolidated earnings, with offshore wind projects in particular running over budget and behind schedule. While management has improved execution materially since the spin, a residual project write-down at any point through 2026-27 would hit earnings even as the gas-turbine and Electrification segments continue to ramp. The wind-business cleanup is the housekeeping item that could surprise to the downside in any given quarter.
Optionality
Four options. First, SMR — small modular reactors are the long-tail nuclear option that GEV has positioned through partnerships; if SMR clears regulatory hurdles in 2027-29, GEV is positioned. Second, European grid build-out — REPowerEU and German Energiewende create a second-engine demand source independent of US AI. Third, services revenue compounding — the installed base produces high-margin recurring service revenue that compounds at 6-8% CAGR through cycle and provides downside cushion. Fourth, capital return acceleration — at FCF of $6.5-7.5 billion and a deleveraging balance sheet, GEV has room to accelerate buybacks or initiate a special dividend; none currently priced. A fifth optionality leg is hydrogen-enabled gas turbines: GEV's HA-class turbines are increasingly designed for hydrogen co-firing, which positions the installed base for decarbonisation retrofits across the next decade — a structural revenue line that would extend the post-AI cycle without depending on new-build demand.
The trade
Entry: $1,000-1,080 — current zone acceptable but watch the technical base; aggressive add below $920. Size: 4-5% portfolio target. Stop: $820 (prior consolidation low / 200-day MA). Catalyst date: Q2 print July 22, 2026; slot reservation milestones quarterly. Trim/exit: trim 25% at $1,350, 50% at $1,500; full exit on hyperscaler capex blink or 110-GW slot target shortfall. Conviction: 9/10.