# AI & Energy: $200B in M&A in the U.S. Electric Sector - Hyperscalers Secure Their Power Supply [ENCADRE titre="Key Takeaways"] [ENCADRE contenu=" - **$200 billion** in mergers and acquisitions in the U.S. electric sector driven by AI energy demands. - **Hyperscalers** (Google, Amazon, Microsoft, Meta) invest heavily to secure their power supply. - Renewable energy and grid modernization become critical priorities. "] [/ENCADRE]

Context

According to the Financial Times (06/29/2026), the volume of mergers and acquisitions in the U.S. electric sector reached a record $200 billion in 2026, driven by hyperscalers (Microsoft, Google, Amazon, Meta) seeking to secure long-term power supply. This trend follows a documented sequence: Exelon CEO’s blackout warning (FT, 06/27/2026), the Moratorium Act proposed by Ocasio-Cortez and Sanders (06/25/2026), and GE Vernova’s turbine order backlog extending to 3–5 years.

Data

• U.S. power sector M&A 2026: $200B – annual record (FT, 06/29/2026)
• IEA Electricity 2025: ~945 TWh global data center consumption in 2026; U.S. share ~380 TWh
• GE Vernova: turbine delivery lead times of 3–5 years (backlogs saturated)
• Exelon CEO: public warning of systemic blackout risk (FT, 06/27/2026)
• Moratorium Act: legislative proposal to cap AI data center power consumption (U.S. Congress, filed 06/25/2026)
• Argentum AI: $4.1B contract for 27,000 GB300 GPUs (06/23/2026) – signal of underlying compute demand

Analysis (Mechanism)

The logic is simple, but its implications are profound. The U.S. power grid was designed for slow-growing industrial and residential demand. AI data center demand, however, is expanding at a pace incommensurate with utilities’ planning cycles—requiring 5–10 years of permitting for new capacity (per IEA Electricity 2025 projections, which document the ramp-up from 2022–2026). Hyperscalers need guaranteed gigawatts within 24–36 months. The only short-term solution: acquisitions. Targets include independent power producers (IPPs), gas-fired plants (dispatchable, reliable), nuclear partners (Constellation Energy for Microsoft), and solar developers with long-term PPAs. M&A is a shortcut to bypass regulatory delays.

This movement creates a scarcity effect on U.S. power assets: mid-cap utility valuations rise mechanically as hyperscalers compete for acquisitions.

Probabilistic Scenarios

• Base Case (50%): Power M&A continues in 2026–2027; mid-cap power assets become premium acquisition targets; capitalization rates decline (energy real assets like 2010s real estate).
• Adverse (30%): Moratorium Act passes → regulatory shock → M&A pause → slowdown in U.S. AI compute deployment → competitive advantage shifts to Europe/Asia for AI infrastructure.
• Bull Case (20%): Power M&A boom attracts SWF and infrastructure PE capital → accelerated grid modernization → U.S. resolves energy constraints faster than expected.

Portfolio Implications

Utilities positioned as M&A targets (Constellation Energy, Vistra, NRG Energy) already trade at an acquisition premium. GE Vernova remains a direct proxy for turbine backlogs—pricing power intact with 3–5 years of visibility. For SMRs: NuScale and X-energy benefit if nuclear becomes the AI energy solution of choice. Data center REITs (Equinix EQIX, Digital Realty DLR) face energy-location risk—assets in blackout-prone zones lose leasing value.

Risks & Blind Spots

• The Moratorium Act remains a political proposal (Sanders/AOC) but materializes a regulatory risk that clouds long-term outlook.
• Permitting delays for new capacity are the real bottleneck—M&A redistributes existing assets but does not create net new capacity in the short term.
• Overpayment risk: hyperscalers, in urgency, may pay excessive premiums that weigh on capital returns.

To Monitor

Congress vote on the Moratorium Act · NERC July report on U.S. grid stability · Q2 earnings for Constellation Energy and GE Vernova · New PPA announcements from Microsoft/Google/Amazon · FERC report on data center interconnections.