top of page

Nuclear Energy: The Boring Stock Powering AI's Electricity Problem

  • The Financial View
  • May 14
  • 4 min read

RESEARCH NOTE — MAY 2026 | THE FINANCIAL VIEW

Nuclear Energy: The Boring Stock Powering AI's Electricity Problem

Artificial intelligence has an electricity problem that nobody in the mainstream conversation is talking about seriously. A single large-scale AI data center draws between 50 and 200 megawatts of continuous power — enough to supply between 40,000 and 160,000 homes. Microsoft, Google, Amazon, and Meta are building dozens of these facilities simultaneously. The International Energy Agency projects that global data center electricity consumption will more than double between 2024 and 2030.

Solar panels generate power when the sun shines. Wind turbines spin when the wind blows. Neither of those conditions holds 24 hours a day, 7 days a week, 365 days a year. AI workloads run continuously. The AI companies know this. Which is exactly why Microsoft signed a 20-year nuclear Power Purchase Agreement. Why Google followed. Why Amazon followed. And why Constellation Energy — the largest nuclear fleet operator in the United States — has become one of the most strategically important infrastructure companies in America.

Why Nuclear Is the Answer

Nuclear power plants run at 90%+ capacity factors, meaning they generate electricity roughly 90% of the time regardless of weather conditions. No other zero-carbon electricity source can match that. Natural gas plants run cleaner than coal but emit CO2. Batteries can store renewable energy but cannot cost-effectively provide baseload power for a major data center campus over decades.

For a hyperscaler committing to net-zero emissions while also guaranteeing power availability for AI infrastructure, nuclear is the only option that checks both boxes simultaneously. This is not an ideological preference. It is engineering reality. The PowerPurchase Agreements being signed are 20 years in length precisely because the hyperscalers are making multi-decade infrastructure commitments. They are not hedging on nuclear power. They are betting the company on it.

Constellation Energy: The Company

Constellation Energy (CEG) is the largest nuclear power generator in the United States. It operates 21 reactors across 12 plants in 8 states, with a total generating capacity of approximately 32 gigawatts. Following its separation from Exelon in February 2022, Constellation is a pure-play clean energy company with nuclear as its core asset.

What makes Constellation structurally compelling — beyond the asset base — is the nature of its revenue. PPA contracts with counterparties like Microsoft, Google, and public utilities lock in electricity prices for decades. The revenue is not tied to the spot price of natural gas or coal. It is not sensitive to weather patterns. It is long-duration, fixed-price cash flow — the kind of revenue profile that, in a rational market, deserves a premium multiple.

Constellation is also the beneficiary of the Inflation Reduction Act's Production Tax Credit for nuclear power, which provides meaningful government support per megawatt-hour generated. That subsidy is additive to the PPA revenue and effectively subsidizes the operating cost of maintaining aging reactors.

The Three Mile Island Restart: A Signal, Not Just a Story

In September 2023, Microsoft signed an agreement to purchase all power output from the Three Mile Island Unit 1 reactor — a reactor that had been decommissioned in 2019. Constellation restarted it specifically for this contract. That is a meaningful signal. When a $3 trillion technology company pays to restart a decommissioned nuclear plant, it is not a PR exercise. It is a statement that they cannot source sufficient baseload zero-carbon power any other way.

The Three Mile Island restart received substantial media coverage as a story. The investment community has been slower to price in what it means for Constellation's long-term contract pipeline. Our view is that this represents the leading edge of a structural shift, not an isolated deal.

Our Position at The Financial View

We initiated our Constellation Energy position in early May 2026 at $312.51 per share, holding 54 shares for a total position value of approximately $16,875. As of this writing, the stock is at $303.63 — we are slightly underwater, with an unrealized loss of approximately $480 (-2.84%).

We are holding. The thesis has not changed. Stock price dips in a structurally sound business are not a reason to exit — they are often an opportunity to add. We are not adding yet because we are preserving dry powder for the ALAB earnings catalyst. Our stop loss is set at $290, which represents a roughly 7% loss from our entry. If the stock breaks below that level, we exit and reassess.

The CEG investment is not a momentum trade. It is a multi-year structural position in a company that sits at the intersection of the nuclear renaissance and the AI power supercycle. We are willing to be patient.

The Risk Factors

Two things break this thesis. The first is government policy reversal — specifically, if nuclear Production Tax Credits from the IRA are repealed or significantly reduced. This risk is real given political volatility, but low probability given that nuclear power enjoys bipartisan support, that it creates domestic manufacturing jobs, and that eliminating the credits would put US AI infrastructure at a disadvantage versus international competitors.

The second risk is hyperscaler demand softening. The 20-year PPAs have exit clauses. If AI capital expenditure slows dramatically — perhaps due to a significant model efficiency breakthrough reducing compute requirements — demand for nuclear baseload power could soften faster than the market expects. We view this as a tail risk, not a base case.

The Bottom Line

Constellation Energy is not a tech stock. It does not have a viral product or a disruptive business model. It operates nuclear reactors built in the 1970s and 1980s. And it has signed contracts that will likely make it one of the most critical infrastructure assets in the United States for the next 20 years. Sometimes the most boring position is the most strategic one.

We own it. We are holding it. And we are watching the hyperscaler PPA pipeline for confirmation that this structural demand shift is only accelerating.

Educational content. Virtual $100K portfolio. Not investment advice. — The Financial View | thefinancialview.org

Comments


Get the next analysis before anyone else.

Free research in your inbox. One click to subscribe.

bottom of page