Two merchant-power giants reported third-quarter results a week apart, and the numbers told the same story twice: earnings far ahead of expectations, guidance raised. Constellation Energy reported on Monday, November 9, with adjusted operating earnings of $5.06 a share against $3.04 a year earlier and a consensus of about $3.90. Vistra had reported on Friday, November 6, with ongoing-operations adjusted EBITDA of $2.38 billion, up 50.4% from a year earlier, and its first-ever 2027 guidance. Between the two reports sat the event that now frames both: on Wednesday, November 4, the day after the midterms, Vistra fell 13.8% and Constellation 6.7%.

The question this edition asks is the one the market has been asking since November 4: what do two strong quarters say about the durability of merchant-power earnings — the hedges, the contracted nuclear megawatts, the guidance — and why did the market sell results this strong? The market was never pricing the quarter. It was pricing the next five years of data-center demand, and the election changed the odds on that demand in ways no earnings release can undo in a week.

Start, as this publication’s earnings-report checklist prescribes, with the two EPS numbers. Constellation’s $5.06 of adjusted operating earnings compares with $3.04 a year earlier, a gain of about 66%, and beats the roughly $3.90 consensus by about 30%. GAAP net income nearly doubled to $1.84 billion, or $5.19 a share, from $930 million, or $2.97 a share. Revenue rose 45.4% to $9.55 billion from $6.57 billion, ahead of the roughly $9.23 billion analysts expected.

But the revenue line needs the honest treatment this publication applies to every acquisition-driven quarter, because the 45% growth rate is mostly arithmetic, not demand. Constellation closed its acquisition of Calpine in January 2026, and Calpine contributed $3.49 billion of operating revenue in the quarter. On a pro forma basis — as if the two companies had already been combined a year earlier — third-quarter 2025 revenue would have been about $9.45 billion, against this quarter’s $9.55 billion.

Constellation: the guidance that did the talking

The guidance is where Constellation’s quarter turns from a beat into a statement. The company raised the midpoint and narrowed its full-year 2026 adjusted operating earnings guidance for the second time this year, to $12.50–$12.80 a share from $11.50–$12.50. The midpoint moved from $12.00 to $12.65, and the new floor sits exactly at the old ceiling. One gap invites caution: the company beat the third-quarter consensus by about $1.16 a share but lifted the full-year midpoint by only 65 cents, which suggests part of the upside was either already inside the prior range or will not repeat in the fourth quarter.

What sits underneath the earnings is the contracting story, the most durable-looking part of the quarter. Constellation has signed more than 8,500 megawatts of long-term power agreements this year, including deals with Amazon and Google, and contracted nuclear megawatts convert high wholesale prices into earnings without the fuel-price exposure that haunts gas-fired peers. In October the company also secured a conditional commitment from the Department of Energy for up to $5 billion in federal financing to fund large-scale nuclear uprates, subject to a final loan agreement.

Vistra: the record quarter the market had already sold

Vistra’s quarter, reported three days before Constellation’s, is the mirror image in structure and the opposite in market timing. Ongoing-operations adjusted EBITDA rose 50.4% to $2.378 billion from $1.581 billion a year earlier, a company record, and GAAP net income rose to $1.14 billion from $652 million. Adjusted earnings were $3.23 a share against a Zacks consensus of $2.94, a beat of about 10%. Revenue rose 28% to $6.35 billion from $4.97 billion. The revenue line carries a footnote: consensus estimates differed by provider — one widely cited tally put the consensus at $6.03 billion, another at $6.85 billion — so whether revenue “beat” depends on which estimate an investor was watching. The EBITDA beat is not in dispute on any tally.

The segment split shows where the money was made, and it is broad rather than concentrated. Texas adjusted EBITDA rose 57% to $1.227 billion from $784 million a year earlier, on higher realized energy and capacity prices and strong fleet performance during record-hot weather. The East segment rose 36% to $979 million from $719 million, Retail contributed $98 million against $37 million a year earlier, and the West added $76 million. Every operating segment grew: this was not one market’s windfall, it was a fleet-wide print.

The net income figure needs the same asterisk the checklist prescribes for every hedge-driven quarter. GAAP net income of $1.14 billion includes a $559 million unrealized gain on hedges expected to settle in future years — nearly half the reported profit, and not cash. The adjusted EBITDA measure strips that out, which is why it is the company’s preferred yardstick and the right one to judge the quarter by.

The 2027 guidance that changed the conversation

Vistra’s most consequential announcement was not a third-quarter figure but a first: 2027 guidance, setting ongoing-operations adjusted EBITDA at $7.7 billion to $8.5 billion and adjusted free cash flow before growth at $4.3 billion to $5.1 billion. The 2027 midpoint of $8.1 billion is about 11% above the 2026 midpoint, built on deals already closed or announced — the Cogentrix acquisition, closed October 30, adding about 5,500 megawatts of gas-fired generation, and the Meta power purchase agreements announced earlier in the year. Management also narrowed 2026 guidance to $7.1 billion to $7.5 billion and added $1.5 billion to its buyback authorization.

The honest caveat is the one the checklist prescribes for every first-time guide. About $700 million of the 2027 midpoint, by management’s own August description, is the contribution of Cogentrix and the Meta agreements — so the underlying growth on the pre-deal base is smaller than the headline suggests. The fourth quarter implied by the narrowed 2026 range is about $1.66 billion of EBITDA at the midpoint, roughly 5% below the fourth quarter of 2025. It is a guide built on acquisitions and contracts rather than on merchant prices staying where they are.

The selloff, and what it was actually about

On Wednesday, November 4, Vistra closed down 13.8% at $120.29, a new 52-week low, and Constellation fell 6.7%. The intraday damage was worse — Vistra traded as much as 15% lower, Constellation as much as 9% — and the selling spread across the trade: NRG fell about 13.5%, GE Vernova about 9.3%. The cause was the midterm elections the day before. Democrats won control of the House and the Senate, and in Texas, Democrat Gina Hinojosa defeated Governor Greg Abbott after campaigning on a moratorium on new data centers. KeyBanc downgraded Vistra the same day, cutting it to Sector Weight from Overweight and removing its price target.

The logic of the selloff ran straight from the ballot box to the order book. Texas accounts for about a fifth of the data-center capacity due to come online in the United States by 2030, and the merchant-power thesis of the last two years has been that data-center demand reprices power in the markets where these generators own plants. A governor who campaigned on pausing new data centers and a Congress that may scrutinize data-center development more closely both point the same way: slower growth in the demand that was supposed to lift merchant earnings for the rest of the decade. The new Congress does not sit until January, and any curbing bill would still need the president’s signature or a two-thirds override — but markets price probabilities, not legislation.

Here is the part the earnings releases expose. Vistra reported that as of October 30 it had hedged approximately 100% of its expected generation volumes for 2026, about 98% for 2027, and about 78% for 2028. Constellation’s earnings are increasingly anchored in long-term agreements — more than 8,500 megawatts signed this year alone — and in a nuclear fleet whose output is contracted years ahead. The election does not change what either company earns in 2026 or 2027. The market sold both stocks anyway, because the stocks had been bought as a way to bet on AI data-center growth. A theme trade falls with the theme, regardless of the contracts.

The recovery, and the valuation gap it left behind

The recovery began almost immediately. Vistra rose 10.5% on Friday, November 6, to $124.80, after its results and the 2027 guidance — still about 11% below its close on November 3. Constellation rose on Monday, November 9, after its report, and by the close on Friday, November 13, the two stocks had clawed back much of the damage: Constellation at $280.20, Vistra at $133.01. The selloff was a week, not a regime — at least so far.

The valuation gap the selloff left behind is stated here as a reported fact with a date. At the November 6 closes — Constellation $249.95, Vistra $124.80 — Vistra traded at about 12.4 times 2027 consensus earnings per share, and Constellation at about 18.8 times, on one provider’s tally. The gap reflects real differences — Constellation’s nuclear fleet and long-term contracts against Vistra’s Texas gas exposure — and different guiding metrics, earnings per share versus EBITDA and free cash flow. A multiple is a price divided by an analyst average, not a fact. The gap is reported because it is what the market did after the election, not because it says which company is the better buy.

The durable-versus-one-off ledger

The durable-versus-one-off ledger is the question the selloff was really asking. Durable, on Constellation’s side: adjusted earnings up 66%, guidance raised for the second time this year, more than 8,500 megawatts of long-term agreements signed in 2026, and a conditional $5 billion federal line for nuclear uprates. Durable, on Vistra’s side: record adjusted EBITDA of $2.38 billion across all four segments, 2027 guidance initiated at $7.7 billion to $8.5 billion, about 98% of 2027 generation hedged, and the Cogentrix fleet closed and contributing.

One-off or timing, on both sides: Constellation’s 45% revenue growth is mostly the Calpine acquisition, and the full-year guidance lift of 65 cents is smaller than the third-quarter beat of about $1.16 — part of the upside does not repeat. Vistra’s GAAP net income includes a $559 million unrealized hedge gain, its revenue beat depends on which consensus tally an investor uses, and its 2027 guidance rests on acquisitions and agreements already announced, not on merchant prices holding. And the political risk is real in a way neither quarter can hedge: a Texas governor who campaigned on a data-center moratorium and a Congress that may slow new development are facts about the future demand these valuations assume, not about current earnings.

What this quarter does not resolve

Three things the two quarters do not settle, and an honest reading should name them. First, the 2028 question. Vistra’s hedges cover about 78% of 2028 generation, and Constellation’s long-term agreements extend further — but the merchant-power thesis is a decade-long story about data-center demand, and the election changed the odds on the years beyond the hedge book. The contracts protect 2026 and 2027. They do not protect the growth the market was pricing for 2029.

Second, the Texas question. Vistra’s Texas segment produced $1.23 billion of the quarter’s $2.38 billion of EBITDA, and Texas is where the political risk is concentrated: a new governor and a paused data-center pipeline. Constellation owns Texas plants through Calpine, so the same risk reaches it, though diluted by a national nuclear fleet. Neither company’s guidance assumes a Texas moratorium. Both companies’ valuations now do, to some degree.

Third, the capacity-price question. Both quarters benefited from high realized energy and capacity prices. Capacity prices reset at auction, and prices that rose sharply can fall. The hedges and contracts blunt that risk for two years. They do not eliminate it.

What to watch

Five markers, in order of consequence. First, the Texas data-center pipeline: whether the new governor’s moratorium becomes policy, and whether ERCOT’s queue — more than 474 gigawatts of requests, about 90% data centers — keeps moving. Second, the next PJM capacity auction: whether the capacity prices behind both quarters’ Eastern earnings are a plateau or a peak. Third, Vistra’s 2027 guidance through the year: a first-time guide with no track record — the quarterly prints are the test. Fourth, Constellation’s uprate financing: whether the conditional $5 billion federal commitment becomes a final loan. Fifth, the fourth-quarter prints: Vistra’s implied fourth quarter is softer than a year ago, and Constellation’s guidance leaves room for an ordinary one.

Two strong quarters, one sharp selloff, a partial recovery: the earnings say the business is durable for two years — hedged, contracted, guided upward. The selloff says the market was never buying two years of earnings — it was buying a decade of data-center demand, and the election priced the risk that the decade arrives more slowly.

None of this is investment advice, and nothing here recommends buying, selling, or holding any security. It is a reading of two quarters’ evidence, with the durable separated from the one-off before anyone calls the selloff a buying opportunity.