The Trump administration is preparing to announce an emergency auction aimed at accelerating the construction of new power plants to support the surging electricity needs of artificial intelligence data centers, according to people familiar with the matter. The plan could ultimately require large technology companies to underwrite as much as $15 billion in new generation capacity, reshaping power markets across the eastern US.

The initiative, expected to be overseen by PJM Interconnection, comes as grid operators warn that explosive growth in AI-driven computing is pushing electricity demand beyond levels anticipated just a few years ago. Amazon, Google, Microsoft and Meta are among the largest drivers, as hyperscale data centers consume round-the-clock power comparable to that of midsize cities.

Under the proposal, PJM would run a special auction allowing data-center operators to secure long-term power supply—potentially through 15-year contracts—by directly financing new power plants. The mechanism is designed to ensure that rising industrial demand does not translate into higher electricity bills for households and small businesses.

“I never want Americans to pay higher electricity bills because of data centers,” Donald Trump said earlier this week, signaling that the administration intends to shift the cost burden toward large corporate users.

A Grid Under Pressure

PJM, which supplies electricity to about 65 million people across 13 states from Illinois to Virginia, has repeatedly warned that capacity margins are tightening. Extreme weather, plant retirements and delays in bringing new generation online have coincided with an unprecedented spike in demand from data centers.

US retail electricity prices rose 7.4% year over year in September to an average of 18.07 cents per kilowatt-hour, the highest level on record. Analysts project that data-center electricity consumption could double by 2030, intensifying strain on grids already operating near their limits.

The emergency auction would prioritize reliable, around-the-clock power sources—particularly natural gas and nuclear—seen as best suited for AI workloads that cannot tolerate interruptions.

Market Implications

Investors are closely watching which power producers stand to benefit if the plan moves forward. Companies with significant exposure to PJM markets and existing generation fleets could see immediate upside from higher capacity payments and new construction opportunities.

Vistra Corp., one of the largest independent power producers in the US, operates a diverse portfolio of gas- and nuclear-fired plants within PJM. The company has already struck agreements to supply electricity to data centers, and additional auctions could accelerate earnings growth tied to AI demand.

Constellation Energy, which owns the nation’s largest nuclear fleet, is also well positioned. Nuclear power’s reliability and carbon-free profile make it attractive for data-center operators seeking long-term supply certainty. The company has recently signed high-profile contracts with technology firms, reinforcing investor expectations of sustained demand.

NRG Energy could benefit from its merchant power model and pipeline of gas-fired projects, while Talen Energy, with a concentrated presence in Pennsylvania, has already demonstrated the viability of dedicated power arrangements for large data-center campuses.

Regulated utilities such as Dominion Energy may also gain indirectly. Faster approvals for new generation could expand rate bases in regions like northern Virginia, home to the world’s largest concentration of data centers.

Broader Consequences

For technology companies, the proposal offers supply security but at a higher cost, potentially compressing margins or forcing expenses to be passed on to customers. For power producers, it represents a rare convergence of political support, structural demand growth and rising electricity prices.

Critics argue the approach could amount to a subsidy for utilities, while supporters say it is necessary to preserve grid reliability and keep AI infrastructure anchored in the US. Environmental groups note that the final impact will depend on whether nuclear displaces fossil fuels or merely supplements them.

If implemented swiftly, the auction could mark one of the most significant interventions in US power markets in decades—reshaping how the costs of the AI boom are allocated between households, utilities and some of the world’s largest corporations.