The Nuclear Crown Jewel of the AI Power Trade — 22 GW of Irreplaceable Baseload, Hyperscaler PPAs, and the Calpine Bet That Changes Everything
Constellation Energy is the largest private-sector power producer in the United States and the nation's largest operator of nuclear power plants. Following the January 2026 acquisition of Calpine Corporation for $16.4 billion, CEG controls approximately 55 GW of generation capacity — roughly 10% of all US clean energy. The fleet spans nuclear (~22 GW), natural gas, geothermal, hydro, wind, and solar across every major US power market.
The core investment thesis is simple: nuclear power is the only zero-carbon, 24/7 baseload electricity source that can meet AI data center demand at scale. Hyperscalers (Microsoft, Meta, Amazon) need clean, reliable power 24 hours a day. Wind and solar are intermittent. Gas has carbon. Nuclear is the answer — and CEG owns more of it than anyone.
| Source | Capacity | % of Fleet | Key Markets |
|---|---|---|---|
| Nuclear | ~22 GW | 40% | IL, PA, NY, MD, NJ, TX |
| Natural Gas | ~28 GW | 51% | TX, CA, PJM, ERCOT (via Calpine) |
| Renewables / Other | ~5 GW | 9% | Geothermal (CA), hydro, wind, solar |
CEG is systematically converting merchant power exposure into contracted hyperscaler revenue. The Microsoft TMI deal (835 MW, 20 years) was the opening salvo. Meta's Clinton nuclear plant extension followed. CyrusOne's 380 MW+ gas co-location deal demonstrates the model works for gas assets too. CEG's strategy: lock in hyperscaler PPAs at premium prices, reducing exposure to volatile wholesale power markets.
Unlike Bloom Energy (where 74% of revenue comes from two Brookfield JVs — related-party circular financing), CEG's customer concentration represents genuine end-user demand from the world's largest technology companies. Microsoft and Meta are buying power they actually need for actual data centers. The deals are arm's-length, with auditable terms and regulatory oversight.
That said, the PPA portfolio is still in early innings. Most of CEG's 55 GW fleet still sells into wholesale markets. The transition from merchant to contracted revenue is the core execution story for the next 3–5 years.
Nuclear fuel supply is concentrated, but not in a single-country-kill-switch way like scandium for Bloom Energy. The US imports ~95% of its uranium, primarily from Canada, Kazakhstan, and Australia. The recent US ban on Russian uranium imports (effective 2028) is accelerating domestic enrichment capacity. CEG has long-term uranium supply contracts and fuel fabrication agreements with multiple suppliers including Cameco (CCJ) and ConverDyn. This is a managed risk, not an existential one.
| Year | Revenue | Rev Growth | Op Income | Op Margin | Adj EPS | Growth |
|---|---|---|---|---|---|---|
| FY2022 | $24.4B | — | $1.0B | 4.1% | $4.24 | — |
| FY2023 | $26.4B | +8.2% | $2.5B | 9.5% | $5.70 | +34% |
| FY2024 | $24.0B | -9.1% | $3.8B | 15.8% | $8.44 | +48% |
| FY2025 | $25.2B | +5.0% | $4.5B | 17.9% | $9.50 | +13% |
| Q1 2026 | $11.1B | +64% | $1.5B | 13.5% | $2.74 | +28% |
| FY2026E | ~$36B | ~43% | — | — | $11–12 | ~20% |
Q1 2026 GAAP EPS was $4.49 (includes mark-to-market gains). Adjusted operating EPS is the key metric at $2.74. Calpine acquisition closed Jan 7, 2026 — Q1 was the first full quarter with Calpine consolidated.
Key trend: CEG has been quietly compounding EPS at 20%+ for several years through operational excellence (nuclear fleet capacity factors consistently above 94%), PTC benefits, and aggressive share buybacks. The Calpine acquisition transforms the revenue base — from $25B to $36B+ — while adding gas-fired generation diversity and geographic expansion into Texas and California.
| Metric | Pre-Calpine (Dec 2025) | Post-Calpine (Mar 2026 / TTM) | Assessment |
|---|---|---|---|
| Cash | $3.64B | $864M | Substantially depleted for acquisition |
| Total Debt | $7.25B | $22.47B | More than tripled |
| Net Debt | ~$3.6B net cash | -$21.6B | Transformation from net cash to net debt |
| Debt-to-Equity | ~0.2x | 0.66x | Still conservative for a utility |
| Interest Coverage | — | 8.04x | Comfortable — no near-term debt stress |
| Book Value | — | $33.8B | P/B of 3.0x is reasonable |
| Current Ratio | — | 1.36x | Adequate liquidity |
The Calpine acquisition tripled CEG's debt from $7.25B to $22.5B and burned through $2.8B in cash. This is the primary reason the stock is down 31% YTD — the market is digesting a much more leveraged company. However, at 0.66x debt-to-equity and 8x interest coverage, the balance sheet is still conservative by utility standards. The key watchpoint: can CEG deleverage through FCF generation while simultaneously funding the TMI restart and dividend growth? At $4.6B annual operating cash flow (TTM), the answer appears to be yes — but integration execution is everything.
| Period | Operating Cash Flow | CapEx | Free Cash Flow | FCF Before Growth* |
|---|---|---|---|---|
| FY2023 | $4.2B | -$2.1B | $2.1B | — |
| FY2024 | $4.8B | -$2.5B | $2.3B | — |
| FY2025 | $5.2B | -$2.8B | $2.4B | $3.8B |
| TTM | $4.6B | -$3.4B | $1.1B | $4.2B |
| FY2026–27E | — | — | — | $8.4B |
| FY2028–29E | — | — | — | $11.5–13B |
* Free Cash Flow Before Growth (FCFbG) is CEG's preferred metric — it excludes growth capex for TMI restart, Calpine integration, and new generation investments. TTM FCF appears depressed at $1.1B due to heavy growth capex ($3.4B). The underlying cash generation is much stronger.
Cash flow reality check: The GAAP FCF of $1.1B and P/FCF of 87x look terrible. But this is a company in a heavy investment cycle — TMI restart ($1.6B), Calpine integration, nuclear uprates, and data center co-location investments. Management's FCFbG metric strips out these growth investments to show the underlying earning power: $4.2B TTM, targeting $8.4B in 2026–2027 and $11.5–$13B in 2028–2029. If management delivers, the FCF yield would expand from 1.1% today to 5%+ by 2028 — with a stock price that would likely re-rate accordingly.
| Metric | Value | Context |
|---|---|---|
| Trailing P/E | 23.4x | Reasonable for 20% EPS growth |
| Forward P/E | 23.6x | In line with utility sector premium |
| PEG Ratio | 1.27x | Below 1.5x — growth at reasonable price |
| P/S (TTM / Forward) | 3.3x / 2.75x | Reasonable for a power producer |
| EV / EBITDA | 14.7x | Attractive vs. peers and growth rate |
| P/B | 3.0x | Reasonable given nuclear asset value |
| P / FCF (GAAP) | 87.2x | Distorted by growth capex cycle |
| Dividend Yield | 0.62% | Low but growing at 10%/year |
| Competitor | Ticker | Nuclear Capacity | vs CEG |
|---|---|---|---|
| Vistra Corp | VST | ~5 GW (Energy Harbor acquisition) | Cheaper (16x fwd P/E), gas-heavy Texas fleet, higher beta AI play |
| Talen Energy | TLN | ~2.5 GW (Susquehanna) | Single-site nuclear operator, AWS PPA, smaller and higher-risk |
| NextEra Energy | NEE | ~3 GW (Turkey Point, St. Lucie) | Renewables giant first, nuclear incidental. Different strategy. |
| Duke Energy | DUK | ~10 GW (regulated) | Regulated utility — no merchant upside, but stable |
| PSEG | PEG | ~3.5 GW (Hope Creek, Salem) | Northeast nuclear; CEG co-owns Salem with PEG |
| GE Vernova | GEV | 0 GW (equipment supplier) | GEV supplies steam turbines to CEG. Not a competitor — a supplier and partner |
Competitive moat: Nuclear power plants are essentially impossible to replicate. Building a new reactor costs $15–30B and takes 10–15 years. CEG's 22 GW nuclear fleet — already built, already operating at 94%+ capacity factors, already licensed — represents literally irreplaceable infrastructure. No competitor can build their way into this position. The only way to compete is to acquire existing plants (as VST did with Energy Harbor) or wait a decade for SMRs. CEG's moat is time.
Constellation Energy completes the AI power infrastructure trilogy. BE is the high-risk disruptor (narrative-driven, supply-chain-dependent). GEV is the industrial compounder (real backlog, real cash flow, real moat). CEG is the asset-backed utility compounder — you're not buying a story, you're buying 22 GW of nuclear power plants that are already built, already running, and increasingly contracted to the world's biggest tech companies.
The YTD underperformance is the opportunity. Down 31% while GEV is up 61% and BE is up 194%. The Calpine debt hangover and TMI restart uncertainty have created a valuation anomaly: a company growing EPS at 20%+ trading at 23x earnings with a 1.27 PEG ratio and an EV/EBITDA of 14.7x. For context, GEV trades at 87x EV/EBITDA (distorted by Wind losses, admittedly). CEG is cheap on an EBITDA basis because the market is discounting the Calpine debt risk.
The key to the trade: You're betting that (a) Calpine integration succeeds and generates the projected FCFbG, (b) TMI restart proceeds on schedule, and (c) hyperscaler PPAs continue converting merchant revenue to contracted revenue. If all three happen, CEG compounds at 20%+ EPS growth for years while the multiple re-rates from 23x to 25–28x — a potential double over 3–4 years. If any one of the three fails, the stock probably treads water or drifts lower.
CEG vs GEV vs BE — picking your risk:
Bottom line: CEG at $274 with a $358 consensus target offers a compelling risk/reward in a sector where most names have already run. The 31% YTD decline reflects market anxiety about Calpine debt, not fundamental deterioration. Q2 earnings on August 6 is the next checkpoint. If management delivers clean integration numbers and raises FCFbG guidance, the stock could rebuild momentum quickly. This is the utility stock for people who don't normally buy utility stocks.
Disclaimer: This report is AI-generated for informational purposes only and does not constitute investment advice. The author may hold positions in securities discussed. Past performance is not indicative of future results. All financial data sourced from publicly available filings, company disclosures, SEC filings, and financial data providers. Always conduct your own due diligence before making investment decisions.