Constellation's New Power Deals Are Piling Up. Here's Why the Stock Isn't Reflecting It Yet.

Constellation Energy’s AI-driven power demand story remains intact, but near-term fundamentals are pressured by leverage and integration costs. After its 2025 $26.6B Calpine acquisition, Q2 total debt rose to $13B (+64% YoY) and EPS fell to $1.42 (-48% YoY), even as revenue grew to $7.5B (+18.6% YoY) and adjusted EPS rose to $2.55 (+33.5% YoY). The company has 920MW of long-duration PPAs lined up (15–20 years, starting 2029–2032), including a 20-year restart PPA with Microsoft for Three Mile Island, but regulatory approvals and grid interconnection delays push cash flows out to late 2027+ (Crane) and beyond—supporting a cautious long-term view despite a 51% drawdown from the 52-week high and a 22% YTD drop.

Analysis

The market is likely underestimating the duration mismatch here: the equity is being asked to fund an asset-heavy expansion while the cash flows it is paying for are mostly 3-7 years out. That means near-term upside is capped by balance-sheet optics and higher-for-longer rates, even if the long-dated contracted pipeline is real. In the next 1-3 months, the stock will likely trade more on refinancing math and regulatory milestones than on the size of the PPA backlog.

The cleaner beneficiary is not the generator but the hyperscaler. MSFT and peers get the strategic option value of de-risked power access without loading their own balance sheets, while CEG absorbs execution, interest-rate, and permitting risk. If grid approvals keep slowing, capital could rotate away from nuclear-like scarcity plays toward lower-duration utility cash flows and infrastructure names; that is a relative loser for CEG multiples and a modest tailwind for regulated power proxies and large banks that finance the capital stack, though fee income is not enough to offset credit risk if leverage stays elevated.

The contrarian miss is that consensus is treating contracted demand as equivalent to de-risked equity value. It is not: long-dated PPAs lock in revenue, but they do not solve common-stock dilution from debt service, project delays, or adverse regulatory remedies. Falsifiers are straightforward: a faster-than-expected approvals cadence, materially lower net leverage by the next quarter, or a sharp drop in long-end rates that expands equity duration multiples.

Key Decisions for Investors

• Short CEG on strength over the next 2-6 weeks; thesis is that the market is still pricing too much of the 2029-2032 pipeline into today’s equity value while leverage remains the binding constraint.
• Pair trade: long XLU / short CEG for the next 1-3 months to express ‘cash-flow now’ versus ‘cash-flow later’ duration risk; this should work if rates stay sticky and regulatory headlines remain slow.
• Use MSFT as the cleaner AI infrastructure beneficiary rather than CEG: accumulate on broad AI pullbacks, because power-securement optionality should improve deployment certainty without adding project-finance risk.
• Watchlist alert on CEG if debt/EBITDA does not improve meaningfully by the next earnings cycle; if leverage stays elevated, expect further multiple compression and consider adding to shorts on any post-approval rally.