To The Moon

The AI boom needs something that doesn’t get nearly as much attention as GPUs:

Electricity. Lots of it.

$Amazon.com(AMZN)$  just signed a 20-year power agreement with $Constellation Energy Corp(CEG)$  that could help expand the Calvert Cliffs nuclear plant in Maryland.

And this isn’t just another corporate electricity contract.

Amazon will support more than $3 billion of investment at the facility, including improvements across the existing 1,790 MW plant and approximately 190 MW of additional generating capacity expected to come online between 2030 and 2032. 

The agreement covers 690 MW of power, including the new capacity.

Why does this matter?

Because the biggest bottleneck in the next phase of AI may not be computing power.

It could be power generation.

Data centres are becoming enormous electricity consumers, and companies such as Amazon, Microsoft and Google are increasingly looking for reliable, around-the-clock power.

Nuclear has an obvious attraction: it can provide large amounts of electricity continuously without relying on weather conditions.

And Amazon isn’t simply buying electricity here.

It’s helping create additional nuclear capacity.

That’s an important distinction.

The 190 MW expansion will feed electricity into the broader PJM grid rather than directly powering an Amazon data centre. The long-term agreement also gives Constellation greater revenue visibility and supports plans to keep Calvert Cliffs operating for another 20 years. 

There’s another interesting angle.

Amazon previously abandoned plans for a data-centre project directly next to the Maryland nuclear plant following community opposition. Instead, it has now shifted toward supporting the plant’s expansion and taking power from the wider grid. 

That could become an increasingly important model.

Rather than building data centres wherever land is available and then trying to find enough electricity afterward, technology companies may increasingly have to think about energy infrastructure first.

For investors, that creates an interesting chain:

AI → Data centres → Electricity demand → Grid investment → Nuclear power

The nuclear renaissance isn’t necessarily just about replacing old reactors.

It could increasingly be about supplying the electricity required for the next generation of computing infrastructure.

Of course, there are still risks.

Nuclear projects can face regulatory delays, huge capital requirements and long development timelines. And a 20-year agreement doesn’t automatically translate into immediate earnings growth for investors.

But Amazon’s latest move highlights something I think the market will be watching more closely:

AI needs power.

And companies that control reliable power generation could become an increasingly important part of the AI infrastructure story.

⚡ The next AI trade might not be a chip company. It might be the company producing the electricity.

What do you think gets the biggest benefit from the AI power boom: nuclear, natural gas, renewables, or the grid infrastructure itself?