The Real AI Bottleneck Isn’t Chips

Not software. Not chips. Electricity.

That’s what this whole AI buildout actually runs on — and the grid is nowhere close to ready. Data center electricity demand is tracking toward 35 GW by 2030, up from 17 GW in 2022. Morgan Stanley puts U.S. data center power needs at 74 GW by 2028, with a projected shortfall of roughly 49 GW in available access. That’s not a rounding error. That’s a structural gap, and it’s opening right now.

Here’s the thing — solar and wind don’t solve this. Data centers need 24/7 baseload power. Nuclear does. Which is why Microsoft signed a 20-year deal to restart Three Mile Island, Meta locked in a 20-year, 1.1 GW agreement starting 2027, and the current administration issued executive orders targeting a quadrupling of U.S. nuclear output by 2050. The policy tailwind is real. The deal flow is real.

Slight tangent, but it matters: most of the nuclear conversation gets hijacked by pre-revenue SMR names — Oklo, NuScale — that are still burning cash and trading well off their highs. CEG is already doing the work at scale.

Constellation Energy (CEG) — The Numbers

Q1 2026 revenue came in at $11.12B against a consensus of $8.71B. Full-year guidance stands at $11–$12 EPS, with free cash flow growth projected through 2029. The stock is trading near $294 — roughly 29% off its 52-week high of $412.58 — with an average analyst price target around $366. P/E sits at ~25.5x, below its 3-year average of 28.8x. Seventeen analysts rate it Buy.

What’s interesting is the setup. You’re buying nuclear baseload power under long-term contracts with the largest tech companies on earth, at a discount to its own historical valuation. That’s a different earnings quality than a standard utility. The Calpine acquisition adds complexity and leverage — that’s the bear case, and it’s legitimate. Regulatory timing on license extensions adds uncertainty. If rates stay elevated, income-seeking capital may not rotate back.

But the gap between where power demand is going and where supply actually sits — that part isn’t slowing down.

For informational and educational purposes only. Not investment advice. Trading involves risk, including loss of principal.