AI supply-chain thesis — mapping bottlenecks, focus companies, and supply-chain exposure for investors.
**Bottleneck theme:** Power & Grid **Focus:** $GEV — GE Vernova Inc. GE Vernova ($GEV) is the pure-play energy company spun out of GE in 2024, and it sells exactly what the AI build-out is short of: **electricity and the equipment that moves it.** Where the nuclear names ($OKLO, $SMR, $CEG, $VST) are the long-dated leg of the AI-power trade, $GEV is the near-term, already-profitable one — it books revenue today on the gas turbines and grid hardware that data centers need this decade. ## Three businesses, one tailwind - **Power** — heavy-duty gas turbines (HA-class), nuclear services, and hydro. Gas is the fastest dispatchable new capacity a hyperscaler campus can secure, and turbine slots are reportedly reserved years out. This segment is the profit engine. - **Electrification** — grid transformers, switchgear, HVDC, and the GridOS software layer. Transformers are a global bottleneck, and every new MW of generation or data-center load needs grid kit to connect it. - **Wind** — onshore (solid) and offshore (the problem child: charges, the Vineyard Wind dispute, $ORSTED.CO / $EQNR Dogger Bank exposure). The bear case lives here. ## Why it's an AI trade Hyperscaler 2026 capex is enormous and power, not chips, is the binding constraint. $GEV is the picks-and-shovels of electrons: it supplies the gas turbines and grid equipment to the utilities and IPPs signing data-center deals — customers like $NEE, $SO, $DUK, $D, $VST, $BRK-B (BHE), $SRE, and $AES — and partners with $NVDA (Omniverse digital twins) and $GOOGL / $AMZN (GridOS) on the AI-meets-grid layer. Its upstream pulls from $HWM (turbine castings), $TPIC (wind blades), $STM (IGBTs for HVDC), $MT (steel), and $DCI (turbine air filtration). ## Financials A genuine, scaling business — not a pre-revenue story. Revenue ~$33B (FY2025) and growing; Q1 2026 delivered ~$8.0B revenue, $254M net income, $0.91 diluted EPS, with gross margin and operating income inflecting positive as Power scales and offshore-wind losses roll off. The balance sheet supports capital return (a ~$1.1B buyback in Q1 2026 plus a dividend), and the multi-year equipment + services backlog gives revenue visibility that pure-play developers can't match. ## Risks - **Offshore wind** — the segment that has generated charges; project disputes (Vineyard Wind) and customer exposure ($ORSTED.CO, $EQNR) are the overhang. - **Data-center demand timing** — management has flagged that some data-center customers "are struggling to get projects across the line"; the order pipeline is strong but lumpy, and the stock is volatile (it fell ~11% in a single month in 2026). - **Valuation & cyclicality** — $GEV has re-rated hard as an AI-power name; it trades on backlog conversion and margin expansion continuing, and it carries the cyclicality of a heavy-equipment OEM. - **Supply-chain constraints** — transformer and turbine lead times cut both ways: pricing power today, execution risk if the chain can't keep up. ## How to play it $GEV is the near-term, cash-generating expression of "AI needs power" — pair it with the nuclear leg ($OKLO, $SMR developers; $CEG, $VST, $TLN operators) for the long-dated side, and with cooling/power-rack names ($VRT) inside the data center. Track the group in the **Power & Grid** basket. ## Related pages - [Oklo ($OKLO) thesis](/thesis/ths_01KTHTC4MXH0ZY09SA559F6TAE-~oklo-stock) — the nuclear leg of the same trade - [SMR Stocks pillar guide](/blog/smr-stocks) - [Data Center Stocks](/blog/data-center-stocks) - [Power & Grid basket](/baskets/power-grid) - [Nuclear Renaissance trend](/trends/mtr_01KKMKE8DG1PX5PFGQQHMGG37N-~nuclear-renaissance) *This is not financial advice — for research and education. Markets are volatile; heavy-equipment cyclicals especially.*
**Bottleneck theme:** Power & Grid **Focus:** $GEV — GE Vernova Inc. GE Vernova ($GEV) is the pure-play energy company spun out of GE in 2024, and it sells exactly what the AI build-out is short of: **electricity and the equipment that moves it.** Where the nuclear names ($OKLO, $SMR, $CEG, $VST) are the long-dated leg of the AI-power trade, $GEV is the near-term, already-profitable one — it books revenue today on the gas turbines and grid hardware that data centers need this decade. ## Three businesses, one tailwind - **Power** — heavy-duty gas turbines (HA-class), nuclear services, and hydro. Gas is the fastest dispatchable new capacity a hyperscaler campus can secure, and turbine slots are reportedly reserved years out. This segment is the profit engine. - **Electrification** — grid transformers, switchgear, HVDC, and the GridOS software layer. Transformers are a global bottleneck, and every new MW of generation or data-center load needs grid kit to connect it. - **Wind** — onshore (solid) and offshore (the problem child: charges, the Vineyard Wind dispute, $ORSTED.CO / $EQNR Dogger Bank exposure). The bear case lives here. ## Why it's an AI trade Hyperscaler 2026 capex is enormous and power, not chips, is the binding constraint. $GEV is the picks-and-shovels of electrons: it supplies the gas turbines and grid equipment to the utilities and IPPs signing data-center deals — customers like $NEE, $SO, $DUK, $D, $VST, $BRK-B (BHE), $SRE, and $AES — and partners with $NVDA (Omniverse digital twins) and $GOOGL / $AMZN (GridOS) on the AI-meets-grid layer. Its upstream pulls from $HWM (turbine castings), $TPIC (wind blades), $STM (IGBTs for HVDC), $MT (steel), and $DCI (turbine air filtration). ## Financials A genuine, scaling business — not a pre-revenue story. Revenue ~$33B (FY2025) and growing; Q1 2026 delivered ~$8.0B revenue, $254M net income, $0.91 diluted EPS, with gross margin and operating income inflecting positive as Power scales and offshore-wind losses roll off. The balance sheet supports capital return (a ~$1.1B buyback in Q1 2026 plus a dividend), and the multi-year equipment + services backlog gives revenue visibility that pure-play developers can't match. ## Risks - **Offshore wind** — the segment that has generated charges; project disputes (Vineyard Wind) and customer exposure ($ORSTED.CO, $EQNR) are the overhang. - **Data-center demand timing** — management has flagged that some data-center customers "are struggling to get projects across the line"; the order pipeline is strong but lumpy, and the stock is volatile (it fell ~11% in a single month in 2026). - **Valuation & cyclicality** — $GEV has re-rated hard as an AI-power name; it trades on backlog conversion and margin expansion continuing, and it carries the cyclicality of a heavy-equipment OEM. - **Supply-chain constraints** — transformer and turbine lead times cut both ways: pricing power today, execution risk if the chain can't keep up. ## How to play it $GEV is the near-term, cash-generating expression of "AI needs power" — pair it with the nuclear leg ($OKLO, $SMR developers; $CEG, $VST, $TLN operators) for the long-dated side, and with cooling/power-rack names ($VRT) inside the data center. Track the group in the **Power & Grid** basket. ## Related pages - [Oklo ($OKLO) thesis](/thesis/ths_01KTHTC4MXH0ZY09SA559F6TAE-~oklo-stock) — the nuclear leg of the same trade - [SMR Stocks pillar guide](/blog/smr-stocks) - [Data Center Stocks](/blog/data-center-stocks) - [Power & Grid basket](/baskets/power-grid) - [Nuclear Renaissance trend](/trends/mtr_01KKMKE8DG1PX5PFGQQHMGG37N-~nuclear-renaissance) *This is not financial advice — for research and education. Markets are volatile; heavy-equipment cyclicals especially.*
The GE Vernova Stock ($GEV): The Power-and-Grid Pick for the AI Electricity Crunch thesis on Macroplane focuses on GE Vernova Inc. (GEV).
Browse all AI supply-chain theses · Macro trends · Industries · Product categories · Baskets