CEG

Good pond, credible fish — CEG is the dominant incumbent in a structurally constrained market (licensed nuclear baseload) where demand is genuinely growing and substitutes don't exist on relevant timelines. The risk is entirely on the price it can extract: at ~23x forward adjusted P/E it needs hyperscaler PPAs at premium terms, and contract-delay risk is documented (34% drawdown from Q1 2026 silence). The nearest peer (Vistra, ~18x P/E) illustrates the multiple compression risk if CEG's clean-nuclear premium erodes.

Constellation Energy is the largest private nuclear power fleet operator in the United States. It owns and operates ~21 GW of nuclear generation across 14 plants, providing 24/7 carbon-free baseload electricity. Following the January 7, 2026 close of its $22B Calpine acquisition, the combined portfolio reaches ~55 GW total generation, adding ~26 GW of gas-fired capacity and making CEG the largest non-utility power producer in the country.

The company sells power through three channels: (1) PJM wholesale energy and capacity markets, (2) retail electricity to commercial and residential customers, and (3) long-term bilateral PPAs with hyperscalers (Microsoft, Meta, CyrusOne) at premium terms. The nuclear fleet is the core asset — already built, already licensed, already grid-connected — with no commercially viable pathway for a competitor to replicate it within a decade.

CEG's strategic agenda for 2026-2027 has four active threads:

1. Crane (TMI Unit 1) restart — a $1.6B investment to restart the shuttered Three Mile Island Unit 1 as the 835 MW Crane Clean Energy Center, underpinned by a 20-year Microsoft PPA. Physical restart targeted for 2027. A FERC waiver (ruling expected June–July 2026) is needed to enable full grid delivery before December 2030 transmission upgrades.

2. LS Power divestiture — selling 5 PJM gas assets (~4.4 GW) to LS Power for ~$5B by the September 4, 2026 DOJ deadline (required as part of the Calpine regulatory resolution). Proceeds strengthen the post-Calpine balance sheet.

3. Gas co-location for AI data centers — leveraging the Calpine gas fleet in ERCOT (Texas) to offer co-located power + land + grid connectivity to hyperscalers. CyrusOne's Freestone and Thad Hill deals (combined >1,100 MW) are the first proof-of-concept, with Freestone energization expected Q4 2026.

4. New hyperscaler PPAs — 147 million MWh of uncontracted clean nuclear generation is the primary commercial pipeline. Management confirmed at Q1 2026 earnings that PPA negotiations are ongoing but renegotiation timelines extended following hyperscaler revisions post-Trump AI executive order.

Revenue (FY, $B):

All figures from SEC EDGAR XBRL (CIK 0001868275).

Key observations:

• OCF was negative for four consecutive years (2021–2024), driven by derivative hedging collateral posting during power price volatility. 2025 was the first strongly positive OCF year (+$4.2B). This history means GAAP OCF can diverge materially from adjusted earnings — context required.
• Revenue dipped from $21.6B (2022) to $18.9B (2024) before recovering to $22.7B (2025) — the 2025 uplift is entirely legacy CEG business; Calpine closed January 7, 2026 and will appear in FY2026 financials.
• Net income declined in 2025 ($2.3B vs $3.7B in 2024), reflecting Calpine transaction costs and integration drag. 2025 operating income also declined ($3.1B vs $4.4B in 2024) for the same reasons.
• 2025 GAAP EPS: ~$6.41 ($2,319M / 362M shares). Management's 2026 adjusted EPS guidance of $11–12/share requires significant non-cash addbacks (Calpine purchase-price amortization, derivative marks). The gap between GAAP (~$6.41) and adjusted ($11–12 guided) is structurally wide post-acquisition.

Balance sheet:

• Cash: $3,641M at FY2025-end (confirmed from fact sheet)
• $5B LS Power divestiture proceeds expected by September 2026 (not yet received as of Q1 2026)
• Investment-grade credit rating; no going-concern language

Share count: 327M (end-2022) → 317M (end-2023) → 313M (end-2024) → 312M (end-2025) → 362M (Q1 2026, post-Calpine 50M share issuance). The pre-Calpine buyback trend is now overwhelmed by Calpine dilution. Net: share count is up ~11% from the pre-Calpine 2025 year-end.

Current metrics:

• Price: $272.65 | Market cap: ~$96.9B
• Trailing GAAP P/E: ~42x ($2,319M net income / 362M shares = $6.41 GAAP EPS)
• Price/OCF (2025): ~23x ($96.9B / $4,237M)
• Forward adjusted P/E: ~23–24x on management's 2026 adjusted EPS guidance of $11–12/share
• Utility peer mean forward P/E: ~19x (Vistra ~18x, PSEG lower); CEG trades at ~21% premium

Fair-value range:

• Floor (no-hyperscaler-premium scenario): Core nuclear/retail earnings ~$9–10/share adjusted at a 20x utility multiple implies ~$180–200. This is the base case if new PPAs are signed at commodity-adjacent terms or continue to be delayed.
• Bull scenario: At $11–12/share adjusted EPS (management guidance), each 10 GW of uncontracted nuclear signed at ~$10/MWh premium adds ~$350–400M EBITDA. Two to three meaningful hyperscaler deals could add $50–100 to the stock. Return toward sell-side consensus ~$380 is achievable within 12–18 months if FERC approves the Crane waiver, one new hyperscaler PPA is signed in 2H 2026, and LS Power divestiture closes on schedule.
• Current $272.65 already embeds some hyperscaler optionality — it is not cheap on core earnings but not at peak-mania levels either.

Fair value range: $180–$380 | Method: Two-scenario multiple expansion — no-PPA floor at 20x core adjusted EPS ($9–10/share) = $180–200; full-optionality scenario at 23–25x on $11/share + incremental PPA EBITDA = $350–380. Current $272.65 sits mid-range, pricing in 1–2 future deals at partial premium.

• Institutional ownership: Dominant — CEG is a NASDAQ large-cap (~$97B) covered by 16+ Wall Street analysts with Buy consensus and ~$380 sell-side price target. Index funds, active managers, and utility-focused funds are the primary holders.
• Insider activity: Not extracted from EDGAR in this analysis; no insider-selling red flags or promotional activity surfaced in any reviewed source.
• Buybacks: $5B authorization in place; $335M deployed in Q1 2026 (1.2M shares at ~$279/share avg). At this pace the authorization takes ~4 years to exhaust — modest relative to the 25M share lockup expiry on June 30, 2026.
• Calpine dilution: 50M shares issued in January 2026 as merger consideration, representing ~16% share count inflation from 312M (end-2025) to 362M (Q1 2026). The net effect of all buybacks since 2022 is now overwhelmed by the Calpine issuance.

CEG is structurally positioned at the only node where AI-power demand meets existing licensed capacity. Three points anchor the bull:

1. Irreplaceable moat. The ~21 GW nuclear fleet is already licensed, already grid-connected, already earning via PJM capacity markets and PTCs. No new-build nuclear is commercially available within 10–15 years. No substitute exists at gigawatt scale for 24/7 carbon-free baseload. Hyperscalers with 2030+ clean-energy commitments have no realistic alternative.

2. Earnings floor is structurally locked. The IRA Section 45U nuclear PTC (up to $15/MWh when power prices fall below ~$44.75/MWh) provides a quantified earnings floor — management estimated $500M in 2028 incremental base revenues. PJM capacity revenues for 2026/27 ($329/MW-day, ~15,500 MW) and 2027/28 ($333/MW-day, already cleared) are locked-in revenue layers on top of energy markets. Even in a scenario with zero new hyperscaler PPAs, the core earnings power ($9–10/share adjusted) supports a $180–200 floor.

3. 34% drawdown may have overshot. The stock fell from ~$412 to ~$273 because one earnings call had no new PPA announcement. The structural drivers (data-center demand, nuclear scarcity, clean-energy mandates) are unchanged. At $272.65, three sequential positive catalysts (FERC approval, one hyperscaler deal, LS Power close) would plausibly support a reversion toward $350–380. The risk/reward is approximately 2:1 asymmetric (upside ~$100 to consensus vs downside ~$70 to floor) — provided IRA credits remain intact.

1. Valuation premium is priced on option value not yet delivered (HIGH severity) Trailing GAAP P/E ~42x; forward adjusted P/E ~23-24x vs peer mean ~19x. The premium is only justified if new hyperscaler PPAs are signed at genuinely above-market terms. If new contracts price at commodity-adjacent rates (PJM energy + capacity), there is no incremental value creation and the stock de-rates toward $210–220 (19x on $11 adjusted EPS). Sell-side consensus at ~$380 reflects bull-case scenario anchoring, not a base-case valuation.

2. Contract delay risk — documented and causal (HIGH severity) The 34% drawdown from October 2025 peak to April 2026 was directly caused by zero new hyperscaler PPAs at the Q1 2026 earnings call. Management disclosed that PPA renegotiations are ongoing because hyperscalers revised their terms following the Trump AI executive order. Direction of renegotiation is toward lower pricing for CEG; hyperscalers hold leverage as large buyers with alternatives. Each quarter of silence is another quarter in which the option-value premium is discounted further.

3. Lockup expiry — 25M shares on June 30, 2026 (HIGH severity, imminent) ~$6.8B of potential selling supply enters the float in approximately four weeks. Q1 2026 buyback pace (~1.2M shares at $335M) absorbs less than 5% of that supply. Management has not committed to accelerated buybacks. The $5B LS Power proceeds (which could fund acceleration) have not yet been received.

4. Crane restart — grid-delivery bottleneck with adverse regulatory risk (HIGH severity) Full 835 MW of the Crane restart requires FERC to approve transfer of Capacity Interconnection Rights from the retired Eddystone plant. The PJM independent market monitor filed opposition on April 21, 2026, arguing the waiver fails FERC standards and imposes costs on other customers. An adverse FERC ruling limits deliverable Crane capacity until the December 2030 transmission upgrade, reducing the economic return on the $1.6B restart investment. FERC ruling expected June–July 2026.

5. Calpine integration complexity (MEDIUM severity) 50M new shares issued (+16% dilution), mandatory $5B LS Power divestiture required by DOJ, purchase-price amortization creates wide GAAP/adjusted EPS gap, gas fleet exposure dilutes the clean nuclear narrative. Q1 2026 adjusted EPS of $2.74 beat consensus by only $0.13 — modest for a transformational $22B deal.

6. Nuclear PTC political risk (MEDIUM severity) IRA Section 45U is the earnings floor. The 2025–2026 GOP reconciliation process is targeting IRA provisions. Nuclear has bipartisan support but is not immune to modification. Management's 2026 guidance explicitly models PTC receipt; a phaseout acceleration or income-threshold change would reduce the earnings floor by an estimated 5–10%.

7. OCF volatility history (MEDIUM severity) Four consecutive years of negative OCF (2021–2024): -$1.3B, -$2.4B, -$5.3B, -$2.5B. The 2023 trough (-$5.3B) was driven by derivative collateral posting. 2025's +$4.2B reversal is real but investors should be aware that stated adjusted earnings can diverge sharply from cash generation in volatile power markets.

• The 34% drawdown precedent is the clearest risk signal: a single earnings call with no new PPA announcement erased ~$45B of market cap. This is not a theoretical risk — it already happened once.
• Revenue arithmetic correction: The $22.7B FY2025 revenue vs $19.0B FY2024 (+$3.7B, +19.5%) is entirely the legacy CEG business improving. Calpine closed January 7, 2026 and does not appear in any 2025 annual figures. Claims that "Calpine doubled revenue" refer to Q1 2026 vs prior year quarterly comparisons, not the annual figures.
• Hyperscaler concentration math: The three signed deals (Microsoft/Crane 835 MW, Meta/Clinton 1,121 MW, CyrusOne ~1,100 MW) total ~3,056 MW contracted to hyperscalers out of ~21,000 MW nuclear fleet — approximately 14–15% contracted. The market premium prices the path from 14% to 30–40%, not just today's 14%.
• OCF reversal is structural, not one-time: 2025's +$4.2B OCF represents derivative hedge settlements flowing through as power prices normalized and the IRA PTC benefit beginning to compound. This should persist into 2026 absent another commodity shock.
• No promotion signals: CEG is a $97B NASDAQ large-cap with 16+ covering analysts. The 'nuclear AI power' narrative creates retail crowding risk rather than promotion risk. The ~34% drawdown itself is evidence the narrative premium partially corrected.

CEG is a genuinely high-quality asset trading at a premium that requires multiple specific things to go right. The structural moat (licensed nuclear fleet, PJM capacity revenues, IRA PTC floor) is real and well-documented. The demand thesis (hyperscaler 24/7 clean power need) is real. What is not real yet is the next tranche of signed PPAs — and the market learned in April 2026 exactly how much that absence costs.

At $272.65, the stock is 34% off peak but still prices in hyperscaler optionality above a no-PPA floor of ~$180–200. The June/July 2026 FERC ruling on the Crane waiver is the first fork in the road: a favorable ruling validates the $1.6B TMI restart investment and removes the transmission bottleneck narrative; an adverse ruling would likely spark another leg down. The June 30 lockup expiry (25M Calpine shares) is the near-term technical overhang.

For long-horizon investors who believe in the data-center power buildout and nuclear's irreplaceable role, the current pullback offers a better entry than October 2025. For anyone requiring near-term catalysts, the next 4–8 weeks are genuinely binary-heavy. The structural earnings floor holds in either case.

Research, not investment advice. Figures sourced from SEC filings and public data; verify before acting.