GE Vernova (GEV)

Gas turbines + grid equipment + wind — the pure-play public bet that AI data-center prime power can ever actually get built.

1. Core Product / Service

GE Vernova (NYSE: GEV) is the energy business spun off from General Electric in April 2024, alongside GE Aerospace. Three reportable segments:

• Power — heavy-duty gas turbines (7HA, 9HA, 6F, 7F frames; aeroderivative LM6000/LM2500/LMS100), steam turbines, hydro turbines, nuclear (GE Hitachi BWRX-300 SMR)
• Electrification — high-voltage transformers, switchgear, grid software (GridOS), HVDC converters, solar inverters
• Wind — onshore wind turbines, offshore wind (post-restructuring)

The AI exposure runs through both Power and Electrification:

• Gas turbines (the large 100-500 MW frame and the smaller aeroderivative class) for prime/utility-scale generation behind hyperscale campuses
• Transformers, switchgear and HVDC equipment sold directly to hyperscalers and to the utilities serving them

GE Vernova is structurally the most-leveraged public equity to the "Bring Your Own Power" thesis: when grid interconnect queues stretch past 5 years, gas turbines from GEV plus reciprocating engines from caterpillar / cummins are how a campus actually turns on.

2. Target Users & Pain Points

• Hyperscalers buying behind-the-meter generation — Microsoft, AWS, Meta, Google directly contracting for gas turbines (skipping the utility) for new campuses
• Independent power producers (IPPs) building gas plants under hyperscaler PPAs — Vistra, Calpine, NRG, Constellation
• Utilities — building gas peakers and reliability assets; buying transformers + HVDC equipment
• Neoclouds — crusoe-energy explicitly named on Q1 2026 earnings call as a customer for gas-turbine-based behind-the-meter power [4]

Pain solved: the only path to multi-hundred-MW prime power that is faster than 5-year utility queues. Gas turbines remain the only commercial-scale dispatchable resource that can be procured (with a lead time) and deployed under hyperscaler timelines. As of Q1 2026, slots through 2030 are largely sold.

3. Competitive Landscape

GEV's edge: largest installed base (>7,000 gas turbines globally), full-stack from generation through transmission to grid software, and the most concentrated pipeline of orders booked through 2030. Disadvantage: physically capacity-constrained — no faster than the rest of the industry on lead times.

4. Unique Observations

• Section 4 focus — 1 MW build-cost share: gas-turbine-based prime power runs ~$600/kW (up from ~$200/kW in 2019, per Wood Mackenzie) [5] → roughly $0.6M/MW of prime capacity. If GEV equipment underwrites both prime gas turbines and transformers/HVDC on a site, GEV's share of a $11-20M/MW AI DC build cost lands in the 5-12% range (much higher when sized for prime rather than backup).
• Q1 2026 is the story: revenue $9.3B (+16% YoY); orders $18.3B (+71% YoY); backlog $163B [1][3]. Adjusted EBITDA $0.9B nearly doubled YoY; margin +390 bps. Full-year 2026 guidance raised to $44.5-45.5B revenue, 12-14% EBITDA margin. 2026 free cash flow guide $6.5-7.5B [1].
• Data-center concentration inside Electrification: in Q1 2026 alone, the Electrification segment booked $2.4B of equipment orders tied to data centers — more than all of 2025 combined [1][7]. Transformers and high-voltage switchgear have moved from "sleepy utility supply" to "scarce AI infrastructure."
• Gas turbine backlog + slot reservations: Q1 2026 ended with 44 GW in firm backlog + 56 GW in slot reservations = 100 GW combined, growing toward a 110 GW year-end target [7]. The 21 GW of new agreements signed in Q1 spans US, Vietnam, Mexico, Brazil, Canada [7]. Big-frame slots through 2030 are effectively sold out.
• Wood Mackenzie data is the bottleneck explanation: gas turbine prices +195% since 2019, reaching ~$600/kW by end of 2027 [5][6]. Large turbines now have ~5-year lead times, smaller units 18-36 months [6]. Turbine orders expected to peak in 2026 with 63 GW of US gas capacity additions through 2030 [6]. This squeeze is structural — adding casting / forging capacity takes years.
• AI revenue share — non-trivial but diluted: Electrification's $2.4B Q1 DC orders implies $10B+ run-rate of DC-attached orders for that segment alone (25-30% of segment revenue). Gas turbines for hyperscaler behind-the-meter are also a meaningful and growing slice. Overall company AI exposure is probably ~15-25% of revenue today and trending toward 30-40% by 2028 as the slot reservations convert.
• Hyperscaler / AI Lab binding: GE Vernova CEO Scott Strazik has publicly discussed volume agreements with hyperscalers stretching to 2035 [4][7]. Named customers include Microsoft, AWS, Meta, and crusoe-energy. This isn't quarter-to-quarter book-to-bill — it's multi-year capacity allocation.
• Token cost chain — the "shovel seller": GE Vernova sits in L1 A.b (prime/utility power) plus L1 A.a (transformers, HVDC, grid equipment). Every GW of hyperscale GPU compute brought online in 2026-2030 will likely consume GEV-supplied gas turbines and transformers somewhere upstream. The 100 GW gas backlog effectively pre-books a fraction of the next 5 years of AI data-center electrification.
• The $200B target moved forward a year: GEV management used the Q1 2026 print to pull forward the prior $200B revenue target to 2029 (from 2030) [3] — reflecting the order-book acceleration.

5. Financials / Funding

• Listed: NYSE: GEV (spun off from GE, April 2024)
• Q1 2026 revenue: $9.3B (+16% YoY) [1]
• Q1 2026 orders: $18.3B (+71% YoY) [1]
• Q1 2026 backlog: $163B (record) [1][3]
• Q1 2026 adjusted EBITDA: $0.9B (nearly 2× YoY); margin 9.6% (+390 bps) [1]
• Q1 2026 Electrification DC orders: $2.4B (>2025 full year) [1][7]
• Gas turbine combined backlog + slot reservations: 100 GW Q1 → 110 GW year-end target [7]
• 2026 revenue guidance: $44.5-45.5B (raised from $44-45B) [1]
• 2026 EBITDA margin guidance: 12-14% (raised from 11-13%) [1]
• 2026 FCF guidance: $6.5-7.5B [1]

6. People & Relationships

• CEO: Scott Strazik (since 2020 GE Power; CEO of GE Vernova from spin-off 2024)
• CFO: Ken Parks
• HQ: Cambridge, Massachusetts
• Founded as spin-off: April 2, 2024 (GE Industries / Power businesses + GE Renewable Energy + GE Digital)
• Sub-brands: GE Vernova Gas Power, Steam Power, Hydro Power, Nuclear, Wind, Grid Solutions, Electrification Software
• Named hyperscaler customers (per Q1 2026): Microsoft, Amazon (AWS), Meta, crusoe-energy — plus IPPs (Vistra, Calpine, NRG, Constellation) building under hyperscaler PPAs
• Strategic context: The most direct public-equity bet on "the only way to power 2026-2030 AI capex is gas + grid equipment, and the supplier is capacity-constrained."