Constellation Is 26% Off, But the Nuclear AI Deal Flow Hasn't Slowed

The pullback made CEG cheaper without changing the core nuclear-AI setup

Constellation is now around $264.2 after a 24.4% one-year decline, even as the alternate energy group gained 21.7% and the broader oil-energy sector rose 21.9%. That makes this less of a routine correction and more of a reset in how the market is valuing the same nuclear-AI story investors were chasing earlier.

Demand is still visible. Reuters reported that rising electricity demand from data centers helped Constellation beat fourth-quarter profit estimates, and the company has kept advancing power arrangements with customers such as CyrusOne, Meta, and Microsoft. The asset base has not changed: Constellation still operates the nation's largest nuclear fleet and says it is positioned to serve rising demand from electrification and the data economy.

That is why the setup matters now. If execution and financing prove manageable, investors are getting the same infrastructure story at a lower entry price.

Demand is showing up in results and in hyperscale deal activity

The bull case starts with a simple question: does Constellation sell something the AI buildout actually needs? The evidence so far says yes.

Data-center demand is already helping reported results

Reuters reported that rising electricity demand from data centers helped Constellation beat fourth-quarter adjusted profit estimates. The same profile noted 23% LTM revenue growth and about $4.6 billion of LTM operating cash flow. That is important because it shows demand is not just a long-term narrative. It is showing up in revenue, earnings, and cash generation.

The Microsoft deal is the clearest proof point

The strongest evidence is the Microsoft arrangement. Constellation said it will sell about 835 megawatts of power to Microsoft and described it as its largest power purchase agreement. The deal also involves restarting Three Mile Island Unit 1. For a buyer of Microsoft's size to commit to nuclear-linked power suggests the demand story is becoming commercial, not purely conceptual.

That does not mean every timeline is risk-free. But it does mean Constellation is no longer asking the market to imagine the demand. In important cases, customers are already signing up for it.

Why the stock sold off: execution and financing are now the debate

The pullback suggests investors still believe in AI-driven power demand. What changed is their willingness to pay a premium before all the execution pieces are proved out.

Calpine, guidance, and shareholder returns are the pressure points

The market is reacting to concern over the Calpine acquisition, including worries about leverage and dilution. That is a real issue for existing holders.

Constellation also guided to $11 per share and $12 per share in 2026, with the midpoint slightly below the $11.60 consensus estimate. At the same time, the company outlined about $3.9 billion of capital spending and raised its buyback authorization to $5 billion. The coexistence of heavier spending and slightly softer guidance helps explain why investors are focusing less on the long-term story and more on how the buildout is funded.

Timing risk is still central to the bear case

Reuters also highlighted agreements with Meta to keep one reactor operating and agreements with Microsoft around Three Mile Island. Those are strategically important, but they do not all translate into the same kind of immediate, incremental power volume.

More broadly, the recent sell-off suggests the market wants proof that projects can advance on schedule and that acquisition-related financing pressure can be contained. If Constellation can do that, today's reset may look like a second-chance entry. If not, the stock could remain trapped between strong strategic demand and slower operational delivery.

What would make the reset look rational over the next one to two years

The next phase is less about pitching the idea and more about proving it can be executed.

Signals that would support the thesis

• More hyperscale power agreements: another major signature deal would reinforce the pattern behind Constellation's largest power purchase agreement with Microsoft.
• Progress on Three Mile Island: clearer movement on the restart of the Three Mile Island plant, including the transmission path needed to make the deal operational.
• A manageable funding story: enough operating cash flow to support the capital spending plan without increasing financing anxiety.
• A steadier fleet backdrop: federal tax credits that continue to support nuclear economics while new demand ramps.

What would weaken it

• Customer timing delays starting to outweigh new hyperscale signings.
• Results that run below the 2026 adjusted-profit range.
• Acquisition-related leverage and dilution concerns that keep squeezing financial flexibility.

If execution improves faster than those risks play out, the de-rating may prove temporary. If not, the market may have been too quick to accept the demand story without waiting for the proof.