2026-10-07 · 2026-10 / week-1
CEG’s Google agreement adds 890 MW, but not a per-share return bridge
CEG’s Google agreement adds 890 MW, but not a per-share return bridge
Summary: Constellation disclosed two Google power agreements: a 20-year PPA tied to 890 MW of new nuclear uprates and a separate 15-year agreement covering 2,700 MW from existing generation. The company says the uprates represent more than $4.3 billion of new investment, but the release gives no power price, operating margin, financing schedule or project return. CEG closed Oct. 6 up 12.25%, then StockAnalysis showed $296.88 after hours at 7:59 p.m. EDT. Even the event-day high is only 4.4% above that after-hours reference, while a return to the prior close is a 9.9% stress. Reject / U.S. long-only no-trade screen. Publication basis: Oct. 7, 2026, 09:02 SGT.
Opportunity Ranking
| Rank | Candidate | Direction | Mispricing | Evidence Freshness | Catalyst Window | Positioning | 10/5 Status | Tradeability | Main Rejection Risk |
|---|---|---|---|---|---|---|---|---|---|
| 1 | Constellation Energy (CEG) | Long | A 3.59-GW headline combines 890 MW of new uprates with 2,700 MW from existing generation; the agreement announces $4.3B of investment but no project-level price or earnings bridge | Oct. 6 issuer agreement; Oct. 6 close and 7:59 p.m. EDT after-hours reference | First uprate expected by 2028; construction, delivery and contract economics remain ahead | Volume was elevated, but no current holder, short-interest, options or flow audit | Reject: 0% reference-base move, -9.9% prior-close stress, 0:1 base/adverse | Nasdaq large-cap common; AH mark only, no current spread/depth/exit audit | Incremental contracted cash flow may be valuable, but its per-share value cannot be derived from disclosed terms |
| 2 | Constellation Brands (STZ) | Long | Fresh Q2 results show 6% sales growth and a 3% comparable EPS increase, but beer depletions fell 0.6%; FY27 comparable EPS range was maintained | Oct. 6 issuer-filed Q2 release after the close; Oct. 6 regular close predates the release | Oct. 7 8:00 a.m. EDT conference call; operating update already published | No current positioning data | Reject: one report and a modest reference mark do not establish a +10% base or bound the downside | NYSE large-cap common; no post-report regular-session acceptance yet | Reported shipment growth and brand share gains may be offset by weak underlying beer depletions and flat FY guidance |
| 3 | Marvell Technology (MRVL) | Long | Oct. 6 Investor Day discussed strategy and growth opportunities, but targets are management projections; stock closed up 5.81% | Oct. 6 official event page and Oct. 6 regular close | Next report date not verified in this run; no new contract identified | No positioning audit | Reject / not selected: no independent valuation bridge from management projections to current equity value | Nasdaq large-cap common; no current execution or exit audit | Customer concentration, execution and valuation remain decisive to translating targets into cash per share |
| Control | SPDR S&P 500 ETF (SPY) | Long | S&P 500 closed at a record, +0.58% on Oct. 6; broad-market strength is context, not a single-name thesis | Oct. 6 regular close | No event-specific catalyst selected | Not assessed | Not selected | Highly liquid fund; liquidity alone is not an edge | A rising index does not bound single-name or macro downside |
Selected opportunity: CEG is the most informative new disagreement to screen, not the best executable opportunity.
Why this one now: Its release quantifies both the physical build and the financing burden: 890 MW of new uprates, a separate 2,700 MW existing-generation contract and more than $4.3 billion of CEG investment. The share price responded sharply, but the release does not disclose the price paid per MWh, expected margin, project schedule by unit, financing mix or return on that capital.
What should surprise the reader: The often-repeated 3.59-GW figure is not 3.59 GW of new generation. Only 890 MW is described as new capacity from uprates; the 2,700-MW agreement is for existing fleet output. That distinction matters when translating a headline into incremental earnings.
Why This Is the Best Opportunity Right Now
The Oct. 6 announcement is a material new evidence boundary from CEG’s Sep. 30 Amazon agreement. That prior contract concerned 690 MW at Calvert Cliffs, including a planned 190-MW uprate expected in 2030–2032. The new Google agreement spans 11 units and names 890 MW of new capacity, with the first uprate expected by 2028; it also includes 2,700 MW from the existing PJM fleet under a separate 15-year supply agreement. The new build is earlier and several times larger than the prior named uprate, but a larger megawatt count is not automatically a larger per-share return.
CEG is more measurable than MRVL’s long-range AI revenue targets, which remain management projections, and more connected to realized operations than the broader SPY control. STZ’s fresh report is useful but is a mixed operating update: comparable EPS increased, while beer depletions were slightly negative and full-year guidance was unchanged. CEG’s contract is the clearest fresh price/positioning/catalyst disagreement because the same-day stock move is large and the new commercial terms expose a specific missing link between contracted power and equity value.
The strongest bull case is real. A 20-year PPA plus a separate 15-year agreement with Google can reduce demand uncertainty for a capital-intensive project. The $4.3 billion investment may create durable incremental generation, and the existing-fleet agreement may help keep plants economically viable. But neither the long term nor the customer name supplies the price, capacity factor, operating costs, taxes, capital structure or cash-flow schedule needed to calculate equity value.
Why This Can Move More Than 5% Soon
The catalyst already produced a greater-than-5% move: CEG closed at $300.40 on Oct. 6, up 12.25% from its Oct. 5 close, on 13.47 million reported shares. The S&P 500 rose 0.58% the same day, so CEG outperformed by roughly 11.7 percentage points. This is a same-session association, not proof that the Google announcement caused the full stock move; other company and market factors may contribute.
The latest after-hours reference retrieved was $296.88 at 7:59 p.m. EDT, 1.17% below the regular close and 10.93% above the prior close. That mark is not regular-session acceptance. A further repricing could follow the first regular session, but the observable contract terms do not yet show whether the market’s new estimate of cash generation is too low or too high.
The cheapest falsification test is to reconcile the eventual project schedule and per-unit approvals with contract pricing, expected incremental generation, operating costs and the $4.3 billion investment. If those disclosures show attractive risk-adjusted returns and the post-event regular market retraces without a change in those economics, a later valuation screen could be rebuilt. Without the economics, the next quote alone cannot prove an edge.
10/5 Asymmetry Gate
| Test | Result |
|---|---|
| Reference basis | $296.88 at 7:59 p.m. EDT / 23:59 UTC, Oct. 6, StockAnalysis after-hours snapshot; reference only |
| Top price reference | $309.80, Oct. 6 regular-session high; +4.4%, historical level rather than a valuation target |
| Highest-probability base reference | $296.88, unchanged from the after-hours snapshot; 0.0% |
| Bottom price stress | $267.62, StockAnalysis Oct. 5 regular close before the Google announcement; -9.9%, not fair value or a loss limit |
| Gross base/adverse ratio | 0:1 |
| Positioning | Reported volume was about 3.85× the Oct. 5 level, but flows, holder identity and short interest are unknown |
| Execution | entry.price = null; execution.can_execute = false; no current post-event regular-session quote, spread, depth, venue-quality or exit-capacity review |
| Classification | Reject / no trade |
The +12.25% event-day gain is already realized from the Oct. 5 close; it is not prospective upside from the $296.88 after-hours reference. The event-day high offers only 4.4% from that mark. The base case cannot reach +10% without an unsupported valuation assumption, while the prior close is a plausible but unbounded retracement stress. This fails the long hurdle even before costs.
What Should Surprise the Reader
The release describes two different economic states. The 890 MW is new generation capacity from uprates, backed by a 20-year PPA and more than $4.3 billion of announced CEG investment. The separate 2,700 MW/15-year agreement concerns energy from existing plants. It is a contract for existing supply, not another 2,700 MW construction program. Combining both contract sizes into a single “new power” number obscures the difference between incremental build and utilization/contracting of existing assets.
The $4.3 billion commitment is also not a profit number. A simple scale ratio is about $4.83 million of announced investment per MW of new capacity ($4.3B ÷ 890 MW). It is not a final unit cost, because the release does not allocate investment across units or schedules, and it says nothing about price, operating expense, financing cost, taxes or return on equity.
The Setup
Fact: CEG and Google announced a 20-year power purchase agreement intended to add 890 MW of new nuclear capacity to PJM through efficiency uprates at 11 CEG units. CEG says the initiative represents more than $4.3 billion of investment; the first uprate is expected by 2028.
Fact: The parties also entered a separate 15-year agreement for 2,700 MW from existing PJM generation. CEG’s Oct. 6 release says the financial structure is intended to ensure existing generation assets continue delivering energy and capacity to PJM. It does not disclose the power price or contract margin.
Fact: In its Aug. 6 Q2 results, CEG reported $2.55 of adjusted operating earnings per share and raised FY2026 adjusted operating earnings guidance to $11.50–$12.50. Its June 30 10-Q reported $1.553 billion of first-half operating cash flow and $2.521 billion of capital expenditures, alongside $1.077 billion of cash and restricted cash. These are company-wide, historical figures and are not a funding bridge for the new Google work.
Inference: The contract may support long-term cash-flow visibility, but the new price increase cannot be validated as undervaluation from megawatts and duration alone. Capacity, energy delivery, cash receipts and shareholder returns are separate steps.
Unknown: Power prices and escalation, per-unit construction schedule and approvals, capital source, incremental operating costs, capacity factors, tax credits, margins, expected return, curtailment terms and the exact timing of cash receipts. I do not have sufficient reliable data to quantify the project’s per-share value accurately.
The Market Price
| Observation | Price / value | Use and limit |
|---|---|---|
| CEG Oct. 6 regular close | $300.40, +12.25%; 13,473,703 shares | StockAnalysis/S&P Global Market Intelligence; event-day context, not an entry |
| CEG Oct. 6 after-hours reference | $296.88 at 7:59 p.m. EDT / 23:59 UTC; -1.17% vs regular close | Latest timestamped AH display retrieved; not an executable quote |
| CEG Oct. 6 regular-session range | $291.00–$309.80 | Provider history; top price reference only |
| CEG Oct. 5 close | $267.62 on StockAnalysis; ChartExchange displays $267.55 | One source discrepancy of $0.07; use StockAnalysis consistently in the simple retracement calculation |
| S&P 500 Oct. 6 close | 7,818.93, +0.58% | AP close summary; broad-market control, not causality proof |
| Announced new capacity / related investment | 890 MW / more than $4.3B | Issuer release; no price-per-MWh or per-share value disclosed |
No current consolidated bid/ask, depth, spread, venue quality, volume quality or realistic exit-size analysis was verified. The AH display and historical session data remain reference inputs only.
The Mispricing
The candidate disagreement is not that Google will consume electricity, or that CEG can sell power. It is whether the market’s +12.25% regular-session revaluation is supported by enough disclosed incremental economics to justify the equity move. The first new 890 MW is a 20-year contracted uprate program with a known aggregate investment headline. The additional 2,700 MW is existing generation under a separate agreement, which can stabilize utilization but does not itself add new nameplate capacity.
Why might the market still be right? A long duration PPA with a high-quality technology counterparty can reduce demand and financing risk, improve fleet utilization and accelerate capital investment. Why might it be wrong? The release omits the price and margin; $4.3 billion of capital has to earn an adequate return, and the agreement’s long life shifts value into distant cash flows. Without debt, taxes, maintenance outages and per-unit timing, investors cannot translate the deal into a defensible net present value per share. The market may already be correctly paying for the option value of contracted growth, or may be overcapitalizing an attractive headline; the public material retrieved does not decide between them.
There is no independent consensus estimate or disclosed implied project value in the sources reviewed. The event-day price/volume response shows attention, not the identity or intent of buyers.
The Positioning
CEG traded 13.47 million shares on Oct. 6 versus 3.50 million on Oct. 5 in the StockAnalysis history, about 3.85× the preceding session. That is a volume observation, not evidence of short covering, index buying or a forced flow. No current short-interest settlement, fund flow, option positioning, borrow data, dealer exposure or holder changes were verified. Positioning score: 2/5.
The Catalyst
| Step | State and timing | Observable test | Failure or delay path |
|---|---|---|---|
| Google 890 MW uprate agreement | Announced Oct. 6; 20-year PPA; investment headline exceeds $4.3B | Read later filings for price terms, unit-by-unit schedule, financing and expected contribution | Delays, approvals, capex overruns, outages or poor realized economics |
| First uprate delivery | Company says first uprate is expected by 2028 | Verify permits, project schedule, commissioning and delivered MW | Schedule slips, output underperforms, or capital rises |
| Existing-fleet 2,700 MW agreement | Separate 15-year supply agreement | Verify contract revenue and contribution relative to prior fleet economics | It may replace other sales or offer little incremental margin |
| Oct. 7 price acceptance | Next regular U.S. session after the announcement day | Compare CEG’s close, range and volume with the after-hours reference and SPY | A retracement could unwind part of the event premium; a continuation still does not prove fair value |
| CEG next earnings report | Issuer date not verified in this run | Reconcile any project, capex, generation and per-share guidance update | No project disclosure leaves the equity bridge unresolved |
Cheapest falsification sequence: Check the next CEG filing and investor presentation for contract pricing or a per-MWh equivalent, and a unit-level capex/schedule bridge. Then compare actual regular-session acceptance and later realized commissioning milestones against the company’s forecast. A repeated capacity headline without pricing or delivery data does not falsify the no-trade decision.
The Payoff
The only expression screened is unlevered CEG common stock. There is no scenario-backed intrinsic target from the disclosed project facts. The simple map below uses observable event-session levels solely as one-day reference stresses: the after-hours mark as the neutral base, Oct. 6’s regular high as a continuation marker, and Oct. 5’s regular close as a retracement stress. None is fair value, a price target or a bounded-loss level.
Price Target and Probability Map
The map runs through the Oct. 7, 2026 regular-session close, the first regular-session test after the Oct. 6 agreement reaction. Probabilities are subjective, low-confidence research weights, not observed frequencies. The base is highest because the available release contains a real long-duration contract but not the economics required to underwrite a further repricing. The top re-tests the session high; the bottom reverts to the prior close. These reference levels are deliberately not represented as fundamental values.
| Scenario | Probability | Target / Level | Return / Payoff | Horizon | Conditions | Evidence Quality |
|---|---|---|---|---|---|---|
| Top Case | 25% | $309.80 | +4.4% | Oct. 7 regular close | Price retests the Oct. 6 regular high as investors continue to capitalize the deal headline | Low; historical intraday level, no valuation bridge |
| Base Case | 55% | $296.88 | 0.0% | Oct. 7 regular close | Price holds near the Oct. 6 7:59 p.m. EDT after-hours reference while the market waits for contract economics | Low; unchanged reference is a neutral scenario assumption |
| Bottom Case | 20% | $267.62 | -9.9% | Oct. 7 regular close | Event premium retraces to the Oct. 5 regular close as investors demand proof of cash returns | Low; pre-event close is a stress, not fair value or a floor |
| Invalidation | n/a | Project price terms, capex and operating schedule demonstrate value materially above the current reference, or the agreement is amended/delayed | Rebuild from disclosed economics and a fresh post-event regular quote | Upon new filing or regular-session price acceptance | Do not treat megawatt totals as earnings or a downside bound | Medium for the decisive evidence required; result unknown |
Probability-weighted expected value: $294.26, about -0.9% gross versus the $296.88 after-hours reference, before carrying and execution costs.
Current market level and timestamp: CEG $296.88 at 7:59 p.m. EDT / 23:59 UTC, Oct. 6; after-hours reference. Latest completed regular close: $300.40 at 4:00 p.m. EDT.
Primary instrument: CEG common stock, screened only.
10/5 favorable base move: 0.0%, below +10%.
10/5 credible adverse move: -9.9% to the historical prior-close stress; losses can exceed it.
10/5 measurement basis: Reference-only; entry.price is null.
10/5 status: Reject / no trade.
Confidence: High that the reference map cannot establish a +10% base case or a 5% downside bound; low on project economics and the subjective scenario probabilities.
Sensitivity and break-even
At 25%/55%/20%, the reference-weighted price is $309.80×25% + $296.88×55% + $267.62×20% = $294.258, or -0.9% before costs. A more bullish 45%/50%/5% split raises weighted value to about $301.23, only +1.5%; the highest-probability base remains flat and the adverse bound is still unsupported. Even a 100% return to $309.80 produces only +4.4% from the after-hours mark. The calculation is a sensitivity to observed historical price levels, not a forecast distribution or proof of fair value.
The Kill Shot
Strongest counterargument: Google is a long-duration, investment-grade technology counterparty, and CEG’s agreement turns a specific load-growth narrative into signed demand. A 20-year contract tied to 890 MW of new production, plus a 15-year agreement for 2,700 MW of existing output, can extend the cash-flow horizon and support capital deployment. The $4.3 billion commitment may be rational if the contract price and generation costs produce attractive returns. The market’s 12.25% reaction may be the start of a justified multi-year revaluation rather than an overreaction.
Why it still fails as a long today: The after-hours reference has already absorbed most of the initial announcement move. Retesting the day’s high adds just 4.4%, and the unchanged base scenario offers no reward. The new project’s price, margin, funding schedule and return on invested capital are undisclosed. The agreement is commercially meaningful, but no public evidence supports a 10% highest-probability price move from the post-event reference with a credible adverse path capped at 5%.
Load-bearing assumption: The Oct. 5 close is a useful downside stress after a failed or revalued contract thesis. It is not a floor. A market-wide risk-off move, lower power prices, higher capex, construction delays or revised earnings guidance could take CEG below it.
What Could Go Wrong
- The 890 MW may deliver meaningful, recurring earnings over 20 years; without price terms, this screen may understate the economic value of contracted nuclear capacity.
- $4.3 billion of investment may produce lower or higher returns than investors expect; the cost headline alone is not a project NPV.
- The 2,700 MW existing-fleet agreement may stabilize generation but may replace other contracts rather than add fully incremental revenue.
- The first uprate is not expected until 2028; construction, permitting, outage timing, component procurement and financing may delay or raise costs.
- A 12.25% close gain and high volume do not prove who owns the shares or that an exit is available at a desired size.
- Nuclear outages, wholesale power prices, fuel costs, tax policy, debt markets and broader market conditions can all move the equity independently of this agreement.
What Would Prove This Wrong
The screen should be rebuilt if CEG discloses the PPA’s pricing/escalators, incremental operating costs, per-unit uprate schedule, funding plan and expected cash contribution in a way that supports a project-level return and an equity-per-share bridge. The first regular-session price acceptance on Oct. 7 would update the reference, not prove the economics. Later filings should verify approvals, construction milestones and delivered MW. A new contract headline without these measures would not be enough.
Risk Audit
The $296.88 mark is an after-hours provider quote last shown at 7:59 p.m. EDT, not an official close or current consolidated quote at publication. No current spread, depth, venue quality, volume quality, order-book size or exit-capacity review was performed. CEG’s event-day high and the Oct. 5 close are historical markers, not support/resistance or risk limits. No price stop can protect against gaps, news, halts or illiquidity. No options, leverage, margin, market order or price-floor logic is supported.
Best Trade Strategy
No trade. Keep entry.price = null and execution.can_execute = false. Reassess after a fresh regular-session quote and evidence that connects the $4.3 billion investment, contract pricing, delivered capacity, debt funding and operating cash flow to per-share value. Require a supported +10% base, no credible adverse path beyond -5%, and at least 2:1 gross reward/adverse risk before any long can qualify. No options, leverage, margin, market orders or price floors.
Sources
| Source | Date / timestamp | Use |
|---|---|---|
| Google–Constellation 890 MW and 2,700 MW agreements | Oct. 6, 2026 | Primary release for 20-year 890 MW uprate PPA, $4.3B investment, first uprate expected by 2028, and separate 15-year 2,700 MW existing-generation agreement |
| CEG–Amazon Calvert Cliffs agreement | Sep. 30, 2026 | Prior PPA mechanism and 190 MW Calvert Cliffs uprate for reuse-boundary comparison |
| CEG Oct. 6 Q2 2026 Form 10-Q | Aug. 6, 2026 | First-half operating cash flow, capital expenditures, cash/restricted cash and capital-intensive context; not a project funding bridge |
| CEG Q2 2026 results | Aug. 6, 2026 | Q2 GAAP/adjusted EPS, FY2026 earnings guide and earlier 920 MW of long-term PPAs |
| CEG Oct. 6 close and after-hours mark | Oct. 6, 2026, 4:00 p.m. and 7:59 p.m. EDT | Regular close, OHLC/volume and separately displayed after-hours reference |
| CEG Oct. 5 regular close | Oct. 5, 2026, 3:59:56 p.m. EDT | Cross-check; provider differs by $0.07 from StockAnalysis Oct. 5 close, preserved rather than averaged |
| S&P 500 Oct. 6 close | Oct. 6, 2026 | Index rose 0.58% to 7,818.93; market control only |
| MRVL official Investor Day page and Q2 2027 issuer release | Oct. 6 / Aug. 27, 2026 | Comparator event date and prior reported growth; no detailed target from the large presentation was used in the selected thesis |
| MRVL Oct. 6 regular close | Oct. 6, 2026, 4:00 p.m. EDT | $287.01, +5.81%; comparator market context only |
| STZ Q2 FY2027 issuer-filed release | Oct. 6, 2026 after close | Comparator’s reported results, beer depletions, brand share, FY guidance and cash-flow context |
| STZ Oct. 6 regular close | Oct. 6, 2026, 4:00 p.m. EDT | $115.67 close before the after-hours release; not post-result regular-session acceptance |
Research Quality Scorecard
| Criterion | Score | Basis |
|---|---|---|
| Market disagreement | 4/5 | Large same-day repricing follows new multiyear contracts, but disclosed deal terms do not quantify shareholder value |
| Evidence base | 4/5 | Fresh issuer agreement plus SEC and price records; contract economics, approvals and project-level returns remain incomplete |
| Positioning and flows | 2/5 | Price and volume observed; direct ownership, short, options and flow evidence absent |
| Catalyst path | 4/5 | Contract signed and first 2028 uprate stated; unit milestones and financial tests still lie ahead |
| Payoff architecture | 2/5 | Reference map has a flat base, <5% top retracement and >5% bottom stress; no derived intrinsic targets |
| Invalidation discipline | 4/5 | Pricing, capex, unit schedule and regular-session acceptance are concrete next tests; historical price levels are not protection |
| Differentiated insight | 5/5 | Separates 890 MW of new generation from 2,700 MW of existing supply and ties the $4.3B build cost to the missing return bridge |
| Client value | 5/5 | Shows how to audit the headline without assuming every contracted MW is new or profitable output |
| Total | 30/40 | Publishable no-trade screen; failed 10/5 economics take precedence |
Bottom Line
CEG has signed a material, long-duration Google arrangement and plans more than $4.3 billion of investment to add 890 MW of nuclear uprates. The separate 2,700-MW agreement concerns existing capacity. The contract may be valuable, but neither the energy price nor the return on new capital is disclosed. After a 12.25% regular-session surge, the $296.88 after-hours reference offers no modeled +10% base; a retracement to the Oct. 5 close is a roughly 9.9% stress. Reject; no trade.
AI Illustration Prompt: Create a precise financial-editorial illustration for The Mispricing Desk: two adjacent contract sheets above a simplified PJM grid. The first sheet reads 890 MW NEW UPRATES / 20 YEARS / >$4.3B CEG INVESTMENT / FIRST EXPECTED 2028; the second reads 2,700 MW EXISTING GENERATION / 15 YEARS. Below, show the bridge still blank: PPA PRICE → OPERATING MARGIN → PROJECT FUNDING → CASH PER SHARE. Put CEG $300.40 REGULAR CLOSE and AFTER-HOURS $296.88 — REFERENCE ONLY on a small market strip, with a separate dotted line to OCT. 5 $267.62 — NOT A FLOOR. Use warm paper, graphite, restrained navy and amber. Do not imply a reactor has already delivered, do not portray every contracted MW as new capacity, and add a subtle readable “The Mispricing Desk” watermark.