Is Constellation Stock a Buy After Jumping 12% on Google's 20-Year Nuclear Deal?

Constellation’s $4.3 billion, 20-year agreement with Alphabet is expected to add 890 MW of nuclear capacity through upgrades at 11 existing units, with the company expecting those uprates to be ready in 2028. Together with a separate $3 billion, 20-year Amazon agreement and a 15-year contract covering 2,700 MW from its existing fleet, the deals strengthen long-term contracted demand as AI data centers seek reliable carbon-free power. First-half revenue rose 44.5% year over year to $18.6 billion and EPS increased 92.7% to $5.88; the company expects more than 20% EPS growth through 2029. Shares are down more than 17% this year, and the article notes elevated valuation and exposure to a potential slowdown in AI spending before upgrade revenue begins.

Analysis

The key valuation question is not whether hyperscalers need firm power, but how much incremental, risk-adjusted cash flow CEG retains after uprate capex, outage timing, and contract terms. Long-duration PPAs reduce merchant-price exposure but can also cap upside if power prices rise; headline contract values are not a substitute for disclosed project returns. The 2028 start date leaves a long execution and discount-rate window, so near-term trading may still follow AI-capex sentiment and rates rather than contracted demand.

Second-order beneficiaries include other dispatchable generators and grid/transmission operators if data-center load tightens PJM capacity and congestion. Conversely, higher system costs could prompt regulators to scrutinize who pays for network upgrades and reliability, slowing project approvals or shifting costs back to generators and large-load customers. Existing nuclear output committed to corporate buyers may also intensify competition for remaining firm supply.

Contrarian view: the market may be pricing the scarcity narrative while underweighting that CEG’s contracts could be less lucrative than merchant exposure, and that uprates still require capital, planned outages, and regulatory execution. The article provides no project-level capex, return, escalation, or termination terms; verify these before underwriting the EPS-growth outlook. Over 1–3 months, watch contract economics, guidance, and rate/AI sentiment; over 6–18 months, execution milestones and PJM supply-demand changes matter more.

Key Decisions for Investors

• Avoid chasing CEG solely on the PPA headlines. Consider a staged, relative-value long CEG / short broad utility exposure only if the entry reflects a discount to its recent relative valuation; size for nuclear execution and factor risk. Falsify the thesis if CEG’s guidance or relative earnings expectations weaken while the utility benchmark holds up.
• Treat the 2028 uprate contribution as a watch item, not a near-term earnings catalyst. Before adding, seek disclosure on project capex, outage schedules, expected returns, and which party bears cost overruns; material slippage or weaker economics would undermine the premium.
• Monitor PJM capacity prices, congestion, and large-load interconnection decisions over the next 6–18 months. A sustained rise in system costs or regulatory intervention that delays data-center connections could impair the scarcity thesis; improving firm-load additions would support it.
• For GOOG and AMZN, regard secured power as a constraint-reduction benefit, not a standalone earnings catalyst: track data-center capex utilization and returns. A slowdown in AI infrastructure spending would weaken future contracting demand and likely pressure CEG’s valuation multiple before the new capacity produces cash flow.