CEG · Constellation Energy Corporation · Stock research

_Research note Constellation Energy Corporation (NASDAQ: CEG). Last close referenced: $252.39 (2026-07-17)._

CEG — Constellation Energy Corporation

Current Thesis

The nuclear-baseload-for-AI story that carried CEG through 2025 has spent 2026 unwinding. The stock shed roughly 35% of its market value in the first half, printed a 52-week low of $228.63 on 2026-07-01, and has since recovered to $252.39 (2026-07-17) still ~15% below where the sell-side consensus cluster sits ($296–$305) and ~10% below the $281 secondary price institutions bought at on 2026-06-01. Two things broke simultaneously. Supply: half the Calpine lock-up shares became saleable on 2026-06-30, layering onto the 11M-share June secondary. Demand: on 2026-07-14 New York imposed the country's first statewide moratorium on data centers drawing 50MW or more, freezing the co-location optionality on CEG's largest state footprint

The narrative leg an investor would be buying here is no longer acceleration. It is stabilization: a merchant-nuclear fleet with a §45U price floor, an $11–$12/sh 2026 guide reaffirmed under Calpine, trading at ~22x while the market re-prices how quickly AI load actually converts to contracted megawatts. That is a value-recovery argument dressed in a momentum theme's clothing, and it does not have a fresh catalyst in front of it until 2026-08-06. The theme leg reads SATURATED and rolling coverage arrived at the highs, the first target cuts have landed, and the demand side just acquired a political constraint that did not exist in June.

Bullish and bearish views on Constellation Energy Corporation

The model's bull view on Constellation Energy Corporation (CEG), in brief: Guide intact through the first full Calpine quarter (2026-05-11): Q1 adjusted EPS $2.74 vs $2.54 consensus, revenue $11.1B vs ~$8.5B estimated, full-year 2026 adjusted operating earnings reaffirmed at $11.00–$12.00/sh. The bear view: New York moratorium, 2026-07-14: Governor Hochul's executive order suspends new and expanded data centers requiring ≥50MW for up to a year while a Generic Environmental Impact Statement is prepared, with DEC barred from issuing discretionary permits not already deemed complete. Both cases follow in full.

Bull Case

• Guide intact through the first full Calpine quarter (2026-05-11): Q1 adjusted EPS $2.74 vs $2.54 consensus, revenue $11.1B vs ~$8.5B estimated, full-year 2026 adjusted operating earnings reaffirmed at $11.00–$12.00/sh. The operating business has not missed.
• §45U nuclear PTC floor (~$43.75/MWh, inflation-adjusted) runs through 2032, capping merchant downside if PJM clearing prices soften the reason a -35% drawdown is a multiple compression rather than an earnings reset.
• Offtake broadened beyond hyperscalers, 2026-06-23: the Walmart PPA from the Dresden Clean Energy Center (Illinois — outside the New York order's reach) demonstrates corporate demand that is not dependent on a single tech counterparty.
• TMI/Crane restart pulled forward, 2026-06-02: the FERC expedited-integration waiver moves the target to 2027 from a prior ~2031 outline, adding contracted capacity in PJM rather than New York.
• Fleet life extended, 2026-06-26: NRC license-renewal applications to run Ginna and Nine Mile Point Unit 1 to 2049 secure roughly two more decades of cash flow regardless of near-term co-location politics.
• Buyback capacity: ~$3.5B of repurchase authority remains after the ~$558M taken down alongside the June secondary a bid that scales as the price falls.
• Valuation reset is real: ~$90.6B market cap at ~21.9x trailing is the cheapest CEG has looked since the Calpine close, against a name where 2027 guidance explicitly excludes future data-center PPAs.

Bear Case

• New York moratorium, 2026-07-14: Governor Hochul's executive order suspends new and expanded data centers requiring ≥50MW for up to a year while a Generic Environmental Impact Statement is prepared, with DEC barred from issuing discretionary permits not already deemed complete. This is the first state-level cap on the load growth the entire thesis rests on, and precedent risk to PJM states is the larger issue.
• Supply overhang is mechanical, not sentiment: half the Calpine lock-up released 2026-06-30, directly ahead of the 2026-07-01 flush to $228.63. The remaining tranche is still ahead.
• Sell-side is walking numbers down: Citi cut its target to $297 from $348 on 2026-07-01 (a ~15% reduction) while holding Neutral, citing model updates after a PJM reliability-risk meeting. Goldman's $305 Neutral (2026-06-18) and Bernstein's $296 Outperform (2026-06-17) were both set roughly 20% above where the stock now trades those numbers are stale and will be revised toward spot, not away from it.
• Mainstream coverage marked the top: Cramer told viewers to buy the dip on 2026-06-15 near the highs; the stock is ~12% lower since.
• Discretionary holders left first: Third Point fully exited in Q1 (13F, 2026-05-15). (Berkshire's 2026-05-15 trim was Constellation Brands / STZ — a different company, frequently misattributed.)
• Credit: the ~$16.4B Calpine close added meaningful leverage with Moody's on review; spread widening beyond 25bps would signal the balance sheet becoming a constraint on the buyback and on further deals.
• PJM interconnection queue is the binding physical constraint, not FERC rulings new load cannot be served faster than the queue clears regardless of how many PPAs get signed.

Setup & Price Structure

Price structure is broken and repairing, not basing from strength. The $281 secondary print (2026-06-01) failed as support and became resistance. The 2026-07-01 capitulation low at $228.63 is the reference low; the ~$252 area is a bounce off it that has not yet been tested. A stock down 35% in six months carries a declining 200-DMA far overhead, so any advance from here is a counter-trend move until proven otherwise the honest characterization is a name attempting to build a right shoulder, with the first real evidence being a weekly close that holds above $281 and turns the failed secondary level back into support.

The beginner-trap read: this is not peak retail sentiment and it is not stretched above its moving averages those risks are behind it. The live trap is the opposite one. A high-quality name 35% off its high, with a reaffirmed guide and a visible PTC floor, is precisely the setup that invites adding on weakness into an unresolved supply overhang and an unresolved regulatory question. There is no structural reason to be early here. The $228.63 low was made 16 trading days ago and the second Calpine lock-up tranche has not cleared.

Catalyst Calendar (next 30 days)

• 2026-08-06 (est., confirmed by consensus calendars) Q2 2026 earnings. Consensus looks for ~+17% YoY quarterly EPS growth. The three gradeable items: whether the $11–$12/sh 2026 guide holds, whether Calpine synergy targets get raised, and any framing of New York exposure.
• ~2026-08-13 (est., 60 days from the 2026-07-14 order) Empire State Development is directed to publish a Community Investment Framework for large data-center projects in New York. First read on how restrictive the eventual permitting regime becomes.
• Undated, Q3 2026 FERC docket ER24-2172 (Amazon/Susquehanna co-location) remains live. This is a binary 15–25% mover in either direction; it has no scheduled date, so it must be watched as a docket, not a calendar entry.
• Ongoing Moody's review of the post-Calpine capital structure. No fixed date.

Elapsed catalysts

• Undated, Q3 2026 second Calpine lock-up tranche release. The first half hit on 2026-06-30 and preceded the low by one session. _(passed 19d ago)_

What Would Change Our Mind

The bearish read fails if the stock reclaims and holds the $281 secondary level on a weekly basis while the August print raises Calpine synergy targets that combination would mean the supply overhang has cleared and the earnings power is compounding faster than the multiple contracted, which is the only sequence that turns a 35% drawdown into a re-rate rather than a range. A new hyperscaler PPA sourced from PJM assets (Illinois, Pennsylvania, Maryland) rather than New York would do similar work by demonstrating the demand leg routes around the moratorium.

Conversely, the constructive case dies on a weekly close below $228, which takes out the capitulation low and confirms the second lock-up tranche is being distributed into a market with no marginal buyer. A second state adopting New York's framework particularly one inside PJM would convert a one-state political event into a sector-wide demand cap and would justify a materially lower multiple on the entire independent-power complex.

Correlation Notes

• Direct peers: VST and NEE were named alongside CEG in the 2026-07-16 coverage of the New York order; the three trade as a single expression of the AI-electricity trade and correlate tightly on regulatory headlines. Any read on CEG that is not confirmed by VST is a single-name story, not a theme.
• Nuclear-specific: TLN (Talen) is the purest merchant-nuclear comparison and drew a bullish initiation on 2026-06-18 near the sector highs. Divergence where TLN holds and CEG does not would isolate the Calpine leverage and lock-up as CEG-specific.
• Upstream equipment: GEV (GE Vernova) is the turbine-side proxy and was initiated bullish on 2026-06-17. Equipment orders lag PPA signings, so GEV weakness would confirm the demand slowdown is real rather than a supply-driven price event in the generators.
• Uranium complex: CCJ and the uranium miners share the nuclear narrative but not the merchant-power P&L; they will not confirm or deny a PJM pricing problem.
• Macro: as a capital-intensive, leveraged merchant generator, CEG carries duration risk long-end yield expansion pressures it independently of the AI narrative, which matters because the Calpine debt load was underwritten in a lower-rate frame.
• Constellation Technology Ventures took an undisclosed equity stake in Blue Energy on 2026-07-16 venture-scale, no near-term P&L relevance, noted only so it is not mistaken for a demand-side datapoint.