Google’s Nuclear Deal Reveals the Scale of AI’s Power Problem

Six Years. Eleven nuclear reactors. Three states. More than $4.3 billion in investment.

That’s what it will take for Constellation Energy (CEG) to squeeze another 890 megawatts of generating capacity out of its nuclear fleet. After all that work, the additional capacity will amount to less than a single gigawatt… Roughly the output of one large nuclear reactor.

Google is helping make it happen through a 20-year power purchase agreement (PPA). And even though the first power increase won’t happen for about two years, this is one of the faster ways to add nuclear power.

These plants already have operating sites and connections to the grid. Google says upgrading them can deliver new capacity years sooner than building a reactor from scratch. Even with that head start, the work will stretch into the next decade.

That tells us these hyperscalers need much more power for their AI data centers than we are currently generating and planning to add. Tech companies can design more powerful chips, develop better models, and commit trillions of dollars to data centers. But those facilities cannot earn the revenue their owners expect until the electricity arrives.

Google’s agreement reveals the scale of that challenge and why supplying power is such an important part of the AI investment story.

Not-So-Short Shortcut

The deal, announced October 6, supports upgrades at 11 nuclear units in Illinois, Pennsylvania, and New Jersey. The industry calls these projects “uprates.” They improve equipment such as turbines, steam generators, and control systems so an existing plant can produce more electricity.

Google’s role is to provide a long-term customer. Its commitment to buy power gives Constellation the revenue certainty needed to invest billions in the work. In return, Google helps secure the dependable electricity its growing business needs.

The companies also signed a separate 15-year agreement covering 2,700 megawatts from existing generation. That helps secure access to power, but it does not add another 2,700 megawatts to the grid.

That brings the total agreement up to 3,590 megawatts with only 890 megawatts of expanded capacity.

As the AI industry grows, signing contracts for existing electricity will only take us so far. We need to produce more of it.

34 Expansions

These reactors supply PJM, the regional grid serving parts of the Midwest and Mid-Atlantic. Its annual forecast projects summer peak demand rising from roughly 160 gigawatts in 2027 to 191 gigawatts in 2031. That’s an increase of nearly 31 gigawatts in four years.

Matching that increase would require roughly 34 expansions the size of Google’s deal.

Of course, other sources of generation will help meet the need. The comparison shows the scale of the task. A major investment backed by one of the world’s largest technology companies addresses only a small share of the expected growth on one regional grid.

The supply problem is already showing up.

PJM’s capacity auction for the 2028–2029 delivery year fell about 6.8 gigawatts short of its reliability requirement. These auctions secure commitments from power resources to meet future demand, including a cushion for extreme weather and plant outages.

That does not mean blackouts are certain. It means the auction failed to secure enough dependable capacity to meet the grid’s planning standard.

PJM’s board has warned that large power users are connecting faster than new generation. It also reported that about 15 gigawatts of generating resources had retired since 2022. The grid must serve new customers while replacing some of the supply it has lost.

The national picture points in the same direction. Lawrence Berkeley National Laboratory’s updated reference forecast projects that data centers could account for roughly 11.8% of U.S. electricity use in 2030. That includes all data centers, with AI helping drive their growth.

This is an enormous new claim on America’s electricity supply. Meeting it will require years of construction.

Recommended Links

The Next AI Winners Will Look Completely Different

Marc Chaikin says a new, volatile phase of the AI bull market is starting, and only 5% of stocks are built to survive it. Tomorrow, October 15, he'll reveal why the AI boom is in crisis – and the breakthrough that can help you potentially double your portfolio amid the uncertainty. Get the details here.

The Surprising TOP FIVE AI Plays for October

TONIGHT, October 14, at 8 p.m. ET… Jeff Brown will discuss his brand-new model portfolio with his TOP FIVE AI plays for the next phase of the AI boom… During an urgent strategy session called The October Million-Dollar Cycle (Click here to RSVP.)

AI Factories Need Power After Sunset

This may lead some to ask, “Why not build something faster?”

We will need a range of solutions. Right now, the power source of choice for data centers needing electricity immediately is natural gas. But this is still a fossil fuel and has some carbon emissions. Plus the lead time for large natural gas turbines is reaching five years.

Solar can supply power during the day, and paired with batteries, it can shift some of that output into later hours.

But AI factories need dependable power across the day and night. Meeting that need requires a mix of generation, storage, and grid equipment. Nuclear is valuable because it can provide large amounts of steady output with low operating emissions.

That helps explain why Google is willing to make a commitment lasting two decades. A larger technology budget cannot make all that physical work happen overnight.

The Value of Scarce Power

As investors, the economics of these agreements are worth scrutinizing.

These PPAs with hyperscalers typically come at a premium to the floor price they would otherwise receive by selling the power to utilities. Microsoft’s deal with Constellation, which will revive the Three Mile Island nuclear plant, came with an agreement that was $40-60 above the going rate.

And Meta’s agreement with Constellation to get power from the Clinton Clean Energy Center came at a $20-25 premium.

Using round numbers, a $50 premium leads to $325 million in annual operating cash flow growth per 1,000 MW. So the Google deal has the potential to increase their free cash flow by more than $1 billion a year once it is in full effect.

As we’ve said, we don’t know the exact details. But hyperscalers are willing to pay a premium for guaranteed access to consistent power to run these AI factories.

The Buildout Behind the AI Boom

The opportunity extends to the businesses that supply and support these power projects.

More generation requires turbines, plant upgrades, maintenance, and fuel. Delivering that power requires transformers, substations, switchgear, and transmission lines.

This is why electricity infrastructure is a long-term investment theme. The growth of AI creates physical needs that must be met before the technology can reach its full commercial potential.

Google’s agreement is a major step forward. Its cost and timeline also show how much more investment lies ahead.

More than $4.3 billion for less than one additional gigawatt. Now consider what it will take to deliver the next several dozen.

We’ll be watching for companies benefiting from this insatiable demand. Investors who pick the right companies will profit immensely in the coming years.

Stay Focused. Stay Invested.

Regards,

Nick Rokke
Senior Analyst, Brownstone Research