The Beat Is All Calpine: Legacy Nuclear Revenues Net of Fuel Ran Flat While PJM Capacity Prices Quintupled — Initiating at Hold

Results vs. Consensus

Q1 2026 Scorecard

Three independent pre-print reads bracketed the adjusted operating EPS consensus at $2.56 to $2.59. We use $2.56, the most widely circulated figure. On the wider of those, the beat is +5.8% rather than +7.0%; the conclusion does not change.

Year-Over-Year Comparison

Sequential Comparison

The sequential comparison is basis-broken and seasonally distorted, and we present it only because a reader will otherwise construct it themselves. Q4 2025 is a standalone Constellation quarter; Q1 2026 includes twelve weeks of Calpine. Q1 is also Constellation's heaviest refueling-outage quarter and its strongest weather-driven demand quarter. No inference about run-rate should be drawn from this table.

Quality of the Beat

• Revenue. Inorganic and mark-to-market, not organic. The $4,334 million of reported revenue growth bridges as follows: $2,395 million from the Calpine reportable segment, $1,602 million from the swing in unrealized hedge gains, $489 million from the four legacy regional segments, less $69 million from Other Power Regions and $83 million from the residual unallocated book. That $489 million of legacy gross revenue growth produced just $10 million of incremental revenue net of fuel. The legacy business grew its gross line by buying and reselling more expensive power, not by earning more on it.
• Margins. The 7.5-point expansion in revenue net of fuel margin is a mix effect, not an efficiency gain. Calpine's segment revenue net of fuel margin of 47.0% and the unallocated hedge mark pulled the consolidated ratio up; within the legacy fleet, Midwest's revenue net of fuel margin fell to 49.3% from 60.5% and Mid-Atlantic's fell to 44.0% from 48.6%. Operating and maintenance grew 15.2% against a fleet that roughly doubled in capacity, which is the one genuinely favorable cost datapoint in the quarter, though it is flattered by a stock-based compensation credit management named as a driver.
• EPS. Adjusted operating earnings grew 44.4% while adjusted operating EPS grew 28.0%. The 16-point gap is the 50 million shares issued to Calpine's owners, which lifted the diluted share count 12.7%. GAAP EPS of $4.49 is not comparable to anything: $2.03 of it is an unrealized fair-value gain and $0.49 is decommissioning-related, offset by $0.44 of acquired-commodity-contract amortization and $0.34 of merger and integration costs. The GAAP effective tax rate rose to 24.9% from an unusually low 14.6%; the company does not disclose a separate tax rate on adjusted operating earnings, so the adjusted-EPS impact cannot be isolated from the filing, but the direction is a headwind and the prior-year rate is the anomaly, not this one.
• Cash. Operating cash flow of $425 million against capital expenditures of $1,275 million means the quarter consumed $850 million of cash before financing, versus $699 million consumed a year ago. Q1 is seasonally the weakest cash quarter for this business and the number carries acquisition-related payables and a $450 million refund of prior-year state-program compensation, but it sits awkwardly beside a free cash flow narrative built on 2028 and 2029.

Segment Performance

Constellation now reports six segments after adding Calpine, and the metric the chief operating decision maker uses is revenues net of fuel, defined as operating revenues less purchased power and fuel expense. That metric, not revenue, is where the quarter is actually decided. The table below is the single most important disclosure in the filing and it appears only in the 10-Q, not in the earnings release or the call deck.

Revenues Net of Fuel by Segment

Segment Revenue and Revenue Net of Fuel Margin

Mid-Atlantic

Gross revenue rose 10.9% and revenue net of fuel rose 0.4%, so the entire increase went to purchased power and fuel, which climbed to $1,035 million from $856 million. Nuclear generation in the region was up 1.1% to 13,326 GWh and purchased power volumes fell 14.6%, so the region generated a larger share of what it sold at a higher unit cost of the balance. The zero-emission credit headwind is concentrated here: New Jersey's program concluded in May 2025, taking its $10.00 per MWh reference price to zero.

Assessment: This is the region with the most direct exposure to a PJM data-center contract and to the Crane restart, and it is producing flat economics today. The bull case for Mid-Atlantic is entirely forward-dated.

Midwest

The most instructive segment of the quarter. Gross revenue rose 23.4% while revenue net of fuel rose 0.5%, and the revenue net of fuel margin compressed by more than eleven points to 49.3%. Nuclear generation fell 2.6% to 22,974 GWh, purchased power volumes fell 14.5%, and purchased power and fuel expense rose 58.5% to $878 million. Illinois zero-emission credit reference prices fell to $1.17 per MWh from $9.38, a decline of 87.5%, and management named lower carbon mitigation credit program revenue as a driver.

Assessment: Midwest is where the "nuclear is levered to power prices" argument breaks down most visibly. ComEd capacity reference prices went up more than ninefold and the segment earned four million dollars more than a year ago. The state credit programs are structured to claw back exactly the upside that rising merchant prices create, and that mechanism is doing what it was designed to do.

New York

The steadiest line in the quarter. Revenue up 1.2%, revenue net of fuel up 2.0%, margin essentially unchanged at 71.9%. Nuclear generation fell 4.2% to 6,014 GWh. New York zero-emission credit reference prices fell 19.2% to $14.76 per MWh, a smaller step-down than Illinois.

Assessment: The most contracted and least optionality-bearing region. It behaves like a regulated asset and should be modelled as one.

ERCOT

The only legacy segment to decline: revenue down 7.0%, revenue net of fuel down 2.3%, with nuclear generation down 7.0% to 2,352 GWh and gas, oil and renewables generation down 11.2%. Margin actually improved to 56.5% from 53.8%, so the decline is volumetric rather than economic. Management characterized the ERCOT forward curve as weak in the near term and undervalued beyond 2029.

Assessment: ERCOT is where Constellation has actually signed co-located data-center deals, including the CyrusOne agreement at Freestone that received PUCT net-metering approval, and where the two new projects came online. The legacy nuclear economics in the region are small and shrinking; the ERCOT story for this name is Calpine's gas fleet and the Powered Land model, not the segment line item.

Other Power Regions

Revenue fell 4.4% while revenue net of fuel rose 37.6% and the margin expanded to 18.0% from 12.5%. Total supply and sales volumes in the region fell 13.6%. This is a lower-margin wholesale and retail book where the improvement is real but the base is small at 5.6% of consolidated revenues net of fuel.

Assessment: A $73 million contribution. Welcome, immaterial to the thesis.

Calpine

Twelve weeks of ownership produced segment revenue of $2,395 million and revenue net of fuel of $1,126 million, a 47.0% margin that sits above the legacy Mid-Atlantic and Midwest regions. Calpine generated 26,497 GWh of gas, oil and renewable output and purchased a further 2,089 GWh. Total revenues attributable to Calpine across all line items, including the portions not allocated to the segment, were $3,136 million.

Assessment: The revenue net of fuel margin is the encouraging datapoint and it validates the retail-plus-dispatchable-gas logic of the transaction. But the 10-Q states plainly that "it is impracticable to determine Calpine's earnings since the Acquisition Date," which means the $2 per share of full-year accretion embedded in guidance cannot be marked to a Q1 actual by anyone, inside or outside the company. That is a defensible accounting position and an uncomfortable position for an investor whose entire earnings beat traces to this segment.

Operating Statistics

Two new operating metrics arrive with the acquired fleet and have no prior-year comparative: an equivalent forced outage factor of 4.5% for the gas, oil and pumped-storage hydro fleet as disclosed in the release, and a forced outage factor of 5.1% for the combined-cycle and cogeneration fleet as described on the call. Those are different fleet definitions of the same idea and both are respectable.

"Operationally, the CCGT and cogeneration fleet had a forced outage factor of 5.1%, meaning our units delivered when called upon nearly 95% of the time."
— Shane Smith, EVP and Chief Financial Officer

Market Reference Prices

These are the prices the company itself publishes as context for its results. Read against the segment table above, they are the argument of this report.

Key Topics & Management Commentary

Overall Management Tone: Deliberately low-energy and administrative, by design. Management opened by noting the prepared remarks would be brief because the substance had been delivered at a business and earnings outlook call six weeks earlier, and the call ran under an hour with seven questioners. Posture on operations and capital allocation was confident and specific; posture on the one variable the Street cares about, PJM contracting, was patient and process-oriented rather than committal, with repeated deferral to regulatory timelines the company does not control. Where management was least convincing was on the bridge from a quintupling of PJM capacity prices to the flat legacy earnings the segment disclosure shows, a bridge nobody was asked to build.

1. The Beat Is Acquisition Arithmetic, and Management Said So

The CFO's opening framing of the quarter was unusually clean and worth quoting at length, because it is the whole quarter in three sentences. He attributed the improvement to Calpine first, then to PJM capacity and a compensation credit, then listed three offsets.

"This is $0.60 per share better than the first quarter of last year and consistent with our expectations. Higher earnings for the quarter can mostly be attributed to the EPS accretion from Calpine. As a reminder, our guidance included approximately $2 per share of accretion for Calpine on a full year basis."
— Shane Smith, EVP and Chief Financial Officer

The offsets he named were more planned refueling outage days, lower zero-emission credit pricing across state programs, and higher cost to serve load associated with a winter storm. None was quantified.

Assessment: Management's own bridge is honest and the segment disclosure corroborates it. The problem is what it implies for the forward: if approximately $2 per share of the $11.50 midpoint is Calpine accretion and the legacy franchise is running flat, then organic growth in 2026 is close to the difference between $11.50 and roughly $11.39, which is the prior year's $9.39 plus $2. That is a thin organic contribution to defend a 26-times multiple.

2. The Nuclear Production Tax Credit Is a Floor, and It Is Currently Out of the Money

The growth pitch leads with the credit. The 10-Q says it produced nothing.

"First and foremost, I want to remind you that our long-term outlook is compelling with a base earnings growth rate that exceeds 20% through 2029, anchored by highly visible drivers that include the nuclear production tax credit, which grows with inflation, long-term contracts with high-quality counterparties and durable customer margins supported by the nation's largest commercial and industrial retail platform."
— Joseph Dominguez, President and Chief Executive Officer

The 10-Q filed the same morning states that for both Q1 2026 and Q1 2025 the company "did not record a material nuclear PTC benefit as the estimate of full year gross receipts exceeded the phase-out for annual gross receipts per MWh for most units." Worse for the quarter, the state programs that pass the credit through cut the other way: the filing discloses a reduction to net operating revenue of approximately $285 million from those pass-through obligations, against an increase of approximately $110 million in the prior-year quarter. The company also refunded or offset approximately $450 million during the quarter relating to credits recorded in 2024, and the associated payable fell to $740 million from $1,190 million.

Assessment: This is the single largest identified swing in the quarter's legacy economics, a $395 million pre-tax move, and it explains most of why the four legacy regions ran flat. It is also the correct way to think about the structure: the production tax credit is a downside hedge that pays when merchant prices are low and disappears when they are high, and the state programs claw back the upside on the same axis. An investor buying Constellation for leverage to rising power prices is buying an asset whose contractual architecture is designed to damp exactly that leverage. Management is not wrong to call the credit a visible driver of the floor. The outlook language does not distinguish a floor from a contributor, and this quarter is the reason that distinction matters.

3. PJM Capacity Upside Is Administratively Capped Through 2029/2030

The 10-Q discloses that in February 2026 PJM filed to extend the RPM capacity market price collar, a cap of approximately $325 per MW-day and a floor of approximately $175 per MW-day, to the 2028/2029 and 2029/2030 Base Residual Auctions, and that FERC accepted the revisions in an April 2026 order, citing reliability concerns and data-center load growth. Constellation's own reference tables show PJM capacity at $269.92 per MW-day this quarter.

Assessment: From $269.92 the administrative headroom to the cap is roughly 20%, and it is locked in for four delivery years. This does not break the thesis, but it does bound one of its two main upside vectors, and it means the capacity re-rate that drove the 2024 and 2025 narrative cannot repeat at anything like the same magnitude. The topic did not come up in prepared remarks or in any of the seven analyst exchanges, which is itself informative about where the Street's attention is.

4. Customer Contracting Has Paused, and Management Confirmed It Twice

The most important disclosure of the call is a negative one. Asked directly whether anything prevents signing a bilateral deal ahead of PJM's reliability backstop procurement, the CEO separated customers into two groups.

"As I mentioned during our last update, some customers have been willing to continue advancing project discussions and agreement negotiations while others have chosen to pause and wait for regulatory clarity."
— Joseph Dominguez, President and Chief Executive Officer

In Q&A he added that the pause is not a stop, and that he expects a rapid resumption once the rules land. He also set an explicit clock: PJM has a proposed timeline to vote and submit to FERC in June, faster than the company had hoped, and the CEO said he expects clarity on all these issues by the end of the year.

Assessment: Nine months of narrative for this stock has been "the hyperscaler contract is coming." What management actually said is that some counterparties stopped negotiating. The honesty is creditable. The consequence is that the highest-value catalyst is now dated no earlier than the second half of 2026 and gated on a regulator, which is precisely the wrong risk profile to pay 26 times forward for.

5. The $338 Million Buyback Was a Post-Quarter Statement About Valuation

No repurchases occurred during Q1. The 10-Q discloses that during April and May 2026, prior to the filing, the company repurchased approximately 1.2 million shares for $338 million inclusive of taxes and transaction costs. The CEO described it in plainer terms.

"Over the past few weeks, we have successfully repurchased approximately 1.2 million shares at an average price of roughly $285 per share for a total of $335 million of purchases. These purchases underscore our commitment to disciplined capital allocation and our confidence in the long-term value of the business. The buyback was an intentional statement from management and our Board that we are excited about the growth opportunities ahead, but that at these prices, we see our stock as a compelling use of our cash."
— Joseph Dominguez, President and Chief Executive Officer

The board approved a $4.4 billion increase to the authorization during 2026, leaving approximately $4.7 billion available.

Assessment: Roughly $285 is 4.9% below the reaction-day close, and it is the most credible valuation signal in the quarter because it is management's own money at a specific price. It also usefully bounds the debate: the board thought the stock was compelling at $285 and chose to say so publicly rather than accumulate quietly. Our reading is that $285 is closer to a floor management will defend than to a level they consider cheap.

6. The June 30 Lock-Up Is a Known Supply Event

Half of the 50 million shares issued to Calpine's former owners come free at the end of June, with the balance a year later.

"Yes. Just to remind folks of the context here, in the consideration for Calpine, we issued 50 million Constellation shares to the owners, 25 million are -- the lockup expires on June 30, 2026, and the remaining 25 million are June 30 of 2027."
— Shane Smith, EVP and Chief Financial Officer

The CFO said the size of the authorization was set partly to retain flexibility "to the extent there could be a transaction of note around the lockup," while declining to characterize the holders' intent.

Assessment: 25 million shares is 6.9% of the 361.2 million outstanding, roughly $7.5 billion at the reaction-day close, against $4.7 billion of remaining buyback authority. The company has pre-positioned for a block. That is competent treasury management, and it is also a reason not to expect the buyback to act as a valuation floor in June: authority earmarked for absorbing a block is not authority available to lean against a selloff.

7. Free Cash Flow Before Growth: The One Genuinely New Disclosure

The quarter's only new financial information was a two-year extension of the free cash flow outlook.

"Over 2028 and 2029, we expect to generate between $11.5 billion and $13 billion of free cash flow before growth. Using the midpoint of that range, that represents approximately a 45% increase relative to the $8.4 billion we expect in '26 and '27."
— Shane Smith, EVP and Chief Financial Officer

The CEO conceded the earlier presentation had been unclear, saying the free cash flow outlook was something "we could have probably done a better job of when we provided the business outlook."

Assessment: A 45% step-up in two-year cash generation is a real number and it is the strongest leg of the bull case, because it is largely independent of new contracts. The caveat is that "free cash flow before growth" is a company-defined measure that excludes growth capital, and this quarter's actual operating cash flow of $425 million against $1,275 million of capital expenditure is a reminder that the reported cash statement and the guidance construct are different objects. We would want two more quarters of conversion before capitalizing the 2028 number.

8. Crane: The Restart Is Not the Question, the Capacity Credit Is

PJM's Phase I system impact study for the Crane Clean Energy Center, the Three Mile Island Unit 1 restart contracted to Microsoft for twenty years, identified contingent transmission upgrades with projected in-service dates extending as late as December 2030. In March 2026 the company asked FERC to let it transfer capacity interconnection rights from Eddystone to Crane, requesting a decision by June 1, 2026.

"Steve, it will come on sooner. I mean, what we're talking about is getting full capacity credit for the asset. So I don't want anybody to be under the misconception that the plant won't start sooner. In terms of getting the full capacity credit, right now, the ball is actually in FERC's court."
— Joseph Dominguez, President and Chief Executive Officer

The general counsel added that the company hopes for a FERC response in the June or July timeframe and that PJM acknowledged the importance of the waiver without taking a substantive position.

Assessment: The distinction management drew is correct and material: energy revenue and the Microsoft power purchase agreement do not wait on the capacity credit. But the requested decision date has already slipped in expectation from June 1 to "June, July," and a 2031 full-deliverability date versus a 2027 one is worth real money in a capacity market that clears above $269 per MW-day. This is a binary regulatory item with a resolution inside the next quarter, which makes it the most tradeable event in the story.

9. Two Projects Delivered, Which Is More Than the Sector Managed

The 105 MW Pastoria Solar Project was commissioned in April, connecting through the interconnection facilities of an existing 750 MW combined-cycle plant and contracted to the California Department of Water Resources, with an 80 MW / 320 MWh battery to follow in spring or summer. The 460 MW Pin Oak Creek peaking facility reached commercial operation in ERCOT on April 30. Both were Calpine projects in flight at the time of the acquisition, a point the CEO made himself.

Assessment: The strategic argument for buying Calpine was never the gas megawatts alone; it was the development and commercial capability. Two assets energized inside four months of closing is early evidence for that argument, and it is the most concrete integration datapoint available given that Calpine's standalone earnings cannot be determined.

10. The Freestone Net-Metering Approval Is the ERCOT Template

The Public Utility Commission of Texas approved, subject to conditions, the net-metering application for co-locating a CyrusOne data center at the Freestone Energy Center, following a 380 MW agreement signed in February with an exclusive arrangement for a further 380 MW in phase two. Construction on the enabling substation is underway with energization targeted for the fourth quarter.

Assessment: Management framed this as "an important signal to the market regarding expectations for colocated projects going forward," and the analogy they are drawing is explicit: Texas Senate Bill 6 was feared as a constraint and instead produced a template once the rules were written, and they expect PJM to follow the same arc. The analogy is reasonable. It is also the second time in this report that the operating progress is in ERCOT while the valuation is being underwritten on PJM.

11. Five Gigawatts Into the Queue, and Nothing Committed

"We have submitted approximately 5,000 megawatts of new capacity resources into PJM's interconnection queue, including unique nuclear uprates, new natural gas generation and new battery storage projects."
— Joseph Dominguez, President and Chief Executive Officer

Pressed on whether the company would commit to new build, management declined pending more detail from PJM on which projects qualify and where they sit in the queue.

Assessment: Queue positions are options, not commitments, and they are cheap to hold. The discipline is appropriate. It also means the 2029 and 2030 capacity additions that the growth algorithm eventually needs have not been sanctioned, and the uprates that are in plan do not contribute earnings until 2030.

12. The Balance Sheet Is a Different Company

Constellation assumed approximately $12.6 billion of Calpine debt, including roughly $5.0 billion of project financings across the Geysers geothermal complex, Calpine Construction Finance Company, a construction revolver, a battery project and two other facilities. It issued $2.75 billion of its own senior unsecured notes in January and used the proceeds to repay $2.5 billion of Calpine corporate term loans immediately, $1.25 billion of secured notes in February and $1.4 billion of unsecured notes in March. A separate exchange offer replaced $2.29 billion of Calpine notes with Constellation paper and stripped substantially all restrictive covenants. Goodwill of $11,107 million was recognized on a $21,835 million purchase and is not expected to be deductible for tax.

Assessment: Net debt of roughly $21.7 billion against $5.4 billion at year-end is a genuine change in the risk profile of the equity, and it arrived alongside a 12.7% increase in the share count. Management reaffirmed a commitment to investment-grade metrics and the credit work in the first quarter was fast and competent. The point is not that leverage is dangerous here; it is that the equity now sits behind materially more debt and roughly $11.5 billion of goodwill, and that the free cash flow story has to service both before it returns anything.

Guidance & Outlook

This was a guidance non-event by construction. The range was set at a business and earnings outlook call on March 31, six weeks before the print, and management noted they were already well into the first quarter when they set it. Affirming a six-week-old range after a $0.18 beat is a quiet way of absorbing the beat rather than passing it through.

Implied ramp to the guide. Q1's $2.74 is 23.8% of the $11.50 midpoint, against 22.8% in the prior year, so the quarter is modestly ahead of a normal seasonal pace. Reaching the midpoint requires $8.76 across the remaining three quarters, versus $7.25 in the same period of 2025, growth of 20.8%. Reaching the low end requires $8.26, growth of 13.9%. Reaching the high end requires $9.26, growth of 27.7%. Given that roughly $2 per share of full-year Calpine accretion is largely mechanical and the legacy fleet is running flat, the low end looks close to assured and the high end looks to require either merchant strength or a contract.

Guidance style. Conservative and process-driven. Asked what would build confidence in the upper half of the range, management would not be drawn.

Q: "And then just given kind of the strong results in the first quarter, right, you guys have mentioned they were relatively in line with your expectations. But historically, we've seen kind of the biggest question mark around the first quarter. So just wanted to get some type of sense, I guess, where -- what would you guys need to see kind of going forward to get kind of more confidence in the upper half of the range and understanding you just spoke a little over a month ago."
— Nicholas Amicucci, Evercore ISI

A: "Look, at least another quarter. Look, we -- when you think about the timing of our business update call, we were well into the first quarter. So we had a pretty good sense of how the quarter was going to come out when we set guidance initially."
— Joseph Dominguez, President and Chief Executive Officer

Analyst Q&A Highlights

Whether ERCOT Forwards Can Be Reconciled With the Data-Center Pipeline

The opening question of the call went straight at the contradiction that defines the sector: forward power curves in ERCOT are weak while the announced data-center pipeline is enormous. Management's answer was that the load is not on the system yet and that the interconnection timing is the swing factor, and then handed to the commercial lead for a quantified version of the argument.

Q: "I was wondering if I could get your latest views on the power market, maybe ERCOT, in particular. Just curious your interpretation and viewpoint here as to the weakness in the forwards even despite some of the very strong data center activity in the pipeline that we're seeing there. What do you make of that and thoughts on the evolution of that market?"
— David Arcaro, Morgan Stanley

A: "So there's been over 400,000 megs of large loads in the queue. Obviously, we don't expect anything near that. But the forward market beyond '29 to us, appears like something that's only expecting 10,000 to 15,000 megs. So if we see numbers like 30,000 megs, we believe that the market will see upward pressure. In the meantime, in the short term, we're not surprised by the weakness, and we've been well hedged and protected against it."
— Andrew Novotny, Senior Vice President, Commercial

Assessment: This is the most useful quantification on the call and it cuts both ways. The claim that the curve prices only 10 to 15 gigawatts of incremental large load against a 400-gigawatt queue is a genuine argument that forwards are too low. But "we've been well hedged and protected against it" is also the admission that the company will not capture the move quickly if it comes, which is exactly the pattern the legacy segment results already show.

Whether a Bilateral Contract Can Be Signed Ahead of PJM's Backstop

A recurring line of questioning tested whether the regulatory process is a genuine blocker or a convenient explanation for a slow contracting quarter. Management separated the two: nothing procedurally prevents a deal, but a subset of customers have chosen to wait.

Q: "Joe, maybe just starting on PJM. You noted that some hyperscaler conversations stopped, some continued. Peers have been a little bit more open to working on deals in parallel with the FERC and PJM process. Is anything preventing having a bilateral deal in hand before the RBP? I mean, do you need to match new capacity plus existing capacity to get a contract?"
— Shahriar Pourreza, Wells Fargo

A: "No. First of all, there's 2 questions there, Shar. Nothing is stopping us from moving forward on a deal now. And I think as I indicated during the last update, we see clients that are interested in doing that. And they figured they'll manage whatever comes out of the regulatory process with the tools that they have or other purchases. For other clients, they kind of want to see what this looks like, what the cost implications are, what our solutions look like, how our solutions pair up with other things we're looking at in the market before they're going to move forward. So I don't think this is a full stop. I do think it is a pause to see what this looks like and then a quick resumption hopefully, of those conversations."
— Joseph Dominguez, President and Chief Executive Officer

Assessment: The question was well aimed, because it puts the regulatory excuse next to peers who are transacting anyway. The answer concedes the point without conceding the framing: deals are available, customers are choosing to wait. That is a demand-side hesitation on price and structure, not a legal impediment, and it is a more uncomfortable read than the regulatory story management prefers.

Crane's Timeline and What Would Pull It Forward

The exchange separated two things the market conflates: when the plant produces power, and when it earns full capacity credit. Only the second is in front of FERC.

Q: "So I guess, first on Crane. Any updates on the time line there? And just what should we be watching for to suggest that maybe it comes on sooner than the 2031 connection?"
— Steven Fleishman, Wolfe Research

A: "No, I just see, we're hoping to get a response back from FERC in the June, July time frame. You also saw that PJM acknowledged the importance of the requested waiver without taking any substantive position otherwise. So everything Joe said, I just want to double down on, and really this is about who bears responsibility for the congestion being ultimately relieved by the RTEP projects and Eddystone does not need those CIRs."
— David Dardis, EVP and General Counsel

Assessment: A clean answer with a dated deliverable, which is rare on this call. The waiver decision lands inside the next reporting period and is the closest thing to a discrete catalyst the story has. Watch it as a binary rather than as a directional read on demand.

Reading the Buyback Against the June Lock-Up

A pointed question asked whether repurchasing shares immediately before a lock-up expiry was a signal about the sellers' intentions. Management declined the invitation and redirected to price.

Q: "Okay. Great. And then other question, just it was good to see the buybacks. We do have this first lockup coming up for the Calpine holders the end of June. Any kind of sense on where their heads are at? And should we read in anything into the fact you were willing to buy stock kind of before that kind of came up?"
— Steven Fleishman, Wolfe Research

A: "Look, I don't think you should read anything into the fact that we bought early. I think I covered that in the prepared remarks. We thought that was a very compelling price to be buying back our shares. We're going to be pretty careful about kind of signaling how different investors may be acting in this space."
— Joseph Dominguez, President and Chief Executive Officer

Assessment: The refusal is appropriate and the follow-on from the finance side was more revealing: the authorization was sized partly to accommodate a possible block around the lock-up. Read together, the buyback is a valuation statement and the authorization is an absorption facility, and investors should not assume the same dollars can do both jobs.

Capital Allocation Against the Illustrative Upside Sensitivity

An analyst pressed on how the roughly $0.50 of upside sensitivity flagged in the outlook interacts with the size of the cash available and whether acquisitions or organic investment are more accretive. The answer was framework-level rather than quantitative.

Q: "There's obviously a substantial amount of cash to allocate $5 billion buyback authorized, $8.4 billion of free cash through '27 and even higher run rate thereafter. How does that kind of tie into the $0.50 of upside sensitivity? And do you anticipate incremental investment opportunity to be more accretive versus the $0.50?"
— Shahriar Pourreza, Wells Fargo

A: "I think, Shar, what I would say about the $0.50 is where we wanted to ensure that there was a range or some flexibility is the nature that those investments could take. To the extent that you're bringing development online, it obviously has a longer period of time until it becomes accretive, whereas if it's M&A, obviously, in the case of Calpine of being of scale and you're adding $2 per share of EPS a year later."
— Shane Smith, EVP and Chief Financial Officer

Assessment: The $0.50 was explicitly described as illustrative and assumption-driven, which is honest and also means it should not be capitalized. The more interesting content is the implicit ranking: another Calpine-scale acquisition is the fast path to accretion, organic development is the slow one, and buybacks sit between. After levering the balance sheet to buy $21.8 billion of assets, a company signalling openness to more scale M&A is a risk factor as much as an opportunity.

Cash Conversion Between EBITDA and Free Cash Flow Through 2029

The extension of the free cash flow outlook to 2029 drew the obvious question about conversion, particularly as Crane and other projects ramp.

Q: "I wanted to ask -- I appreciate the free cash flow clarity out to '29. I mean, could you maybe just talk a little bit about the cash conversion between EBITDA and free cash through '29? And as you get some of these projects up and running, like Crane is going to ramp and a few other things in the back end of the plan, how do you kind of think about cash conversion?"
— Nicholas Campanella, Barclays

A: "Nick, it's Shane. I mean it's not going to change materially from what you've seen historically with regard to the nature that most of the cash contribution is from the nuclear fleet. So if you think about the appropriate assumptions around cash tax, maintenance CapEx, how you're accounting for fuel, the conversion won't look significantly different than historical."
— Shane Smith, EVP and Chief Financial Officer

Assessment: Two things are worth noting. First, that most cash contribution still comes from the nuclear fleet, after acquiring 23 gigawatts of gas, which is a useful reminder of where the economics actually sit. Second, the assertion that conversion is unchanged is a claim about an acquired business that the company simultaneously says it cannot separately measure. Modellers should treat the conversion assumption as unverified for at least another two quarters.

When the Nuclear Uprates Start Paying

The final substantive exchange produced the most concrete forward statement on the call, and it was a deflating one.

Q: "Just wanted to get a sense just on the near-term kind of nuclear uprates, how near term are those? And when are we kind of -- what's kind of baked into the 2029 assumptions?"
— Nicholas Amicucci, Evercore ISI

A: "None of the uprates. The capital is out the door, but nothing is showing up as EPS accretion in '29. 2030 is the earliest that they'll add to EPS."
— Shane Smith, EVP and Chief Financial Officer

Assessment: Uprates are the cheapest incremental clean megawatts in the country and they are frequently cited as an underappreciated lever. This answer removes them from the 2029 earnings picture entirely, with the capital spent and nothing to show through the guidance horizon. It is a clean, credible answer, and it should shrink the uprate line in anyone's 2029 model to zero.

What They're NOT Saying

Everything below is disclosed in the Form 10-Q filed the same morning as the release. None of it was raised in prepared remarks or in any of the seven analyst exchanges. The gap between what the filing says and what the call discussed is the most striking feature of this quarter.

1. The $285 million revenue reduction from nuclear credit pass-throughs. The call named "lower ZEC pricing across state programs" as one of three unquantified offsets. The 10-Q quantifies it: a reduction to net operating revenue of approximately $285 million this quarter against an increase of approximately $110 million a year ago, a $395 million pre-tax swing, plus approximately $450 million refunded or offset during the quarter. This is the largest single identified driver of the flat legacy result and it received one clause.
2. That the nuclear production tax credit contributed nothing. The 10-Q states no material credit benefit was recorded in either Q1 2026 or Q1 2025 because gross receipts exceeded the phase-out for most units. The outlook language presents the credit as an anchor of visible growth. Both are true, and only one was said out loud.
3. That legacy regional revenues net of fuel were flat. The segment table exists only in the 10-Q. Nobody asked why four nuclear regions produced $2,284 million against $2,274 million while PJM capacity reference prices rose more than fourfold, and management did not volunteer the bridge.
4. The PJM capacity price collar. FERC accepted the extension of a roughly $175 to $325 per MW-day collar through the 2029/2030 auction in April 2026, a month before this call. It bounds the largest upside vector in the sector narrative and was not mentioned once.
5. The $5,735 million held-for-sale block and its September 4, 2026 deadline. Six generating assets must be under definitive agreement within 240 days of closing. The five PJM assets are contracted to LS Power for $5.0 billion subject to DOJ and FERC review; the ERCOT plant has no announced buyer, and the filing says only that the company is "taking steps to divest" it. Two of the assets being sold sit inside project financings that may require lender consent to release as collateral. Zero questions.
6. That Calpine's earnings contribution is not determinable. The 10-Q says it is "impracticable to determine Calpine's earnings since the Acquisition Date." The approximately $2 per share of full-year accretion is the load-bearing number in the guide, it was restated on the call, and there is no Q1 marker against it.
7. The cost of the winter storm. Named as an offset in prepared remarks and referenced in the 10-Q only as "a significant weather event," with no dollar figure anywhere in either document. For a quarter whose legacy result was flat, an unsized cost item that management chose to call out is a number worth having.
8. Zion Energy Center's 2030 exposure. The 10-Q discloses that Illinois legislation could require the Calpine-acquired Zion Energy Center to shut down by January 2030 absent relief. A newly acquired gas asset with a legislated retirement risk inside the guidance horizon did not come up.

Market Reaction

• Pre-print setup: Closed at $303.63 on Friday, May 8, down 14.1% year to date against an S&P 500 up 8.1%, up 11.9% over twelve months, and up 6.0% over the trailing thirty days. The 52-week closing range entering the print was $247.06 to $403.95, so the shares sat 24.8% below the high. The independent power complex had been sold hard in the two sessions before the print: Constellation fell 3.6% and 2.5%, and peers fell in the same window.
• Reaction session (before the open reporter, same day): Opened at $315.30, a gap of 3.8%, and traded as high as $317.74, up 4.6%. It then fell all session to a low of $288.83, down 4.9%, and closed at $299.69, down 1.3% or $3.94. The peak-to-close give-back was 5.7%.
• Volume: 6.3 million shares against a thirty-day average of 2.8 million, 2.2 times normal.
• Relative: The S&P 500 rose 0.2% and the utilities sector ETF rose 0.9% on the session. Among independent power peers, Vistra rose 2.9%, NRG fell 0.6%, and Talen fell 0.8%. The fade was specific to Constellation, not a sector event.

The shape of the session is the analysis. A 7% earnings beat and an affirmed guide gapped the stock up nearly 4%, and by the close the entire move plus a little more had been surrendered on more than double normal volume, while the closest comparable in the group finished up 2.9%. That is not a macro tape and it is not a rejection of the print. It is a market that opened on the headline and closed on the content.

The content problem is that nothing in the quarter moved the variables the equity is priced on. Guidance was six weeks old and unchanged. The beat traced to an acquisition that closed four months earlier and whose accretion was already in the number. The one new disclosure, the 2028 and 2029 free cash flow range, is three years out and defined on a company-specific basis. And the single item investors have been waiting on since the second half of last year, a PJM hyperscaler offtake, went backwards: management confirmed that some counterparties have paused negotiations pending rules that a regulator will write in June at the earliest. A beat that changes nothing about the only open question is a beat you sell.

It is also worth noting what the fade was not. There was no guidance cut, no adverse surprise in the filing that the market could have read intraday, and no downgrade in the session. The stock simply found no buyer above the pre-print level once the call ended at midday, which is a positioning statement about a name that entered the print down 14.1% on the year and still trades at 26.1 times the guidance midpoint.

Street Perspective

Debate: Is Calpine an Earnings Machine or an Expensive Diversification?

Bull view: The bull case on the Street holds that Calpine bought exactly what Constellation lacked: dispatchable capacity to firm the nuclear fleet, the largest competitive retail platform in the country, and a development organization that has already energized two projects since closing. A 47.0% revenue net of fuel margin in the first partial quarter is above the legacy Mid-Atlantic and Midwest regions, and approximately $2 per share of annual accretion against a $21.8 billion price is a mid-teens earnings yield before synergies.

Bear view: The bear camp contends that the company paid $21.8 billion for assets carrying $11.1 billion of goodwill, issued 6.9% dilution at a price that has since fallen, took net debt from $5.4 billion to $21.7 billion, and now owns a gas fleet with a legislated retirement risk in Illinois and a mandated $5.7 billion divestiture on a September deadline. The accretion claim is unauditable by the company's own admission.

Our take: The strategic logic is sound and the operating evidence so far is encouraging. The valuation question is separate and unresolved: an acquisition that adds $2 per share to a $9.39 base is roughly 21% accretion for 12.7% dilution and a tripling of net debt, which is accretive but not obviously value-creating at that leverage. We are prepared to be convinced. Two quarters of segment reporting with a determinable earnings contribution would do it.

Debate: Does AI Power Demand Reach the Income Statement, and When?

Bull view: Hyperscaler capital spending for 2026 is running nearly 75% above last year and continues to be revised up, the PJM interconnection queue holds hundreds of gigawatts of large-load requests, and Constellation owns the scarcest asset in that market. Once PJM writes the rules, the argument goes, contracting resumes quickly and premium-priced long-term offtakes reprice the whole fleet.

Bear view: The bear camp points at this quarter's segment table. Power and capacity prices already moved enormously and the nuclear franchise earned four tenths of a percent more. Hedging, state credit clawbacks, and the production tax credit phase-out are structural dampeners, not timing effects, and FERC has now capped PJM capacity at roughly $325 per MW-day through 2029/2030. Demand growth may be real and still not accrue to this equity at the speed the multiple implies.

Our take: The bears have the better of it on the near term and the bulls on the long term, and the price is closer to the bull case. The mechanism that converts AI demand into Constellation earnings is a signed bilateral contract, not a spot price, and there is not one yet. Until there is, the segment table is the evidence and the narrative is a forecast.

Debate: Is $285 a Floor or a Waypoint?

Bull view: Management put $338 million of shareholder money to work at roughly $285 in the weeks before the print, having explicitly told the board and the market that the stock was compelling there, with $4.7 billion of authority remaining. A company with visible free cash flow stepping to $11.5 to $13.0 billion across 2028 and 2029 and a buyback that size does not have much downside from a 14% year-to-date drawdown.

Bear view: The skeptics note that 25 million shares come free of lock-up on June 30, roughly $7.5 billion at current prices against $4.7 billion of authority, and that the company has signalled it would use the authorization to absorb a block rather than defend a level. A known supply overhang landing seven weeks after the print is a poor setup for a name whose catalyst is dated to the second half.

Our take: Both are right, and the sequencing favors the bears in the near term. Buyback authority spent absorbing a block is not authority spent supporting the price on a bad tape. We would treat roughly $285 as the level management has revealed a preference for rather than as a floor they are obliged to hold, and we would rather own the shares after the lock-up event than before it.

Model Framework & Valuation

This is an initiation, so there is no prior model to revise. The table sets out the assumptions we are carrying and what would move each one.

Valuation. At the reaction-day close of $299.69 and 361.2 million shares outstanding, the market capitalization is roughly $108.2 billion; adding net debt of roughly $21.7 billion gives an enterprise value near $129.9 billion, before any adjustment for the $19.4 billion nuclear decommissioning trust and the $12.4 billion asset retirement obligation, which broadly offset one another. The shares trade at 26.1 times the $11.50 guidance midpoint, 27.2 times the low end and 25.0 times the high end. Against our estimates that is 25.8 times 2026 and 22.9 times 2027. Annualizing the quarter's $155 million of dividends paid gives a yield of roughly 0.6%, so total return here is a growth-and-multiple argument, not an income one.

Fair value framework. We anchor on 22 to 25 times our 2027 estimate of $13.10, a range of $288 to $328 with a midpoint near $308, roughly 3% above the close. Applying management's own greater-than-20% base algorithm to the guidance midpoint instead would produce a 2027 figure near $13.80 and shift the range up by about 5%, which is the size of the prize if the contracting story resumes on management's timeline. Neither construction produces a return that justifies taking regulatory timing risk. The stock is fairly priced for what is signed and cheap only for what is not.

Thesis Scorecard: Establishing Coverage

This is first coverage, so the scorecard below establishes the pillars we will grade in subsequent quarters rather than scoring a standing thesis. Status tags reflect where each pillar sits after this quarter's print and call.

Overall: The bull pillars are intact and the bear pillars are the ones this quarter moved. Nothing in the print threatens the 2026 guide. What the print did was demonstrate that the legacy franchise does not convert power-price strength into earnings on the timeline the market has been assuming, and that the highest-value catalyst has slipped to the second half at the earliest.

Action: Initiate at Hold with a fair value range of $288 to $328. We would upgrade on a signed PJM offtake of scale, on a favorable FERC waiver that pulls Crane's full capacity credit into 2027, or on the shares trading toward the level management itself bought at. We would downgrade on a guidance cut, on the September 4 divestiture deadline passing without a definitive agreement on the ERCOT asset, or on evidence that the acquired retail book's margins are not durable once purchase accounting normalizes.