Vistra's Best Quarter Ever Lands in Its Worst Tape

Record $2.4B EBITDA, a freshly minted $24-a-share 2030 target — and a stock 43% off its peak, because the market is now pricing Texas politics rather than Vistra's cash flow.

VST · Earnings Call · 2026-11-06

The print nobody wanted to pay for

Vistra's third quarter was, by its own accounting, the best in company history. Adjusted EBITDA of $2.378 billion, up more than 50% year over year; a record commercial availability figure approaching 95% across the thermal fleet; and guidance narrowed with a higher midpoint on both EBITDA ($7.1–7.5B) and free cash flow before growth ($4.3–4.7B). Management introduced 2027 at $7.7–8.5B of EBITDA and $4.3–5.1B of free cash flow, the first full-year frame that includes the Cogentrix close and the Meta PPA. Then Kris Moldovan put a number on the endgame: adjusted free cash flow per share rising from just under $10.50 in 2025 to roughly $14 in 2027 and about $24 by 2030. Meanwhile the tape says something else entirely. Total revenue of $4.4B, up just 4% year over year against that 50%-plus EBITDA jump tells you where the earnings came from: realized prices more than 10% higher per megawatt hour, higher capacity revenue, and dispatch optimization — not a bigger book of business. And the equity sits 42.7% below its September 2025 peak of $217.92, down 17.4% in the last ninety days alone, with a second, fresher drawdown of 26.1% from a July high of $168.98. A $41.9 billion market cap, and a share price going one way while the income statement goes the other.

A call that opened on an election

Jim Burke did not start with megawatts. He started with the midterms: “the election results do not change our fundamental view of the opportunities Vistra has” — James Burke, Chief Executive Officer · 2026-11-06. That sentence is the whole report in miniature — the results changed nothing about the business, but everything about how the business is being valued. What is genuinely new in this quarter's keyword set is political vocabulary that was absent from the prior four quarters' top-30 lists: Governor elect, legislature, guardrails, and — raised only by an analyst, then swatted away — moratorium. Contrast that with August, when the same word was pre-emptively denied: “there isn't a moratorium at this point in time” — Jim Burke, President and Chief Executive Officer · 2026-08-07. By November, asked directly whether Texas could impose a data center moratorium next year, Burke did not argue the mechanics — he argued the politics: “I just don't entertain that, David” — James Burke, Chief Executive Officer · 2026-11-06. Just as telling is the refrain. In August, on ERCOT demand: “From a long-term fundamentals of the business, we haven't actually changed our forecast for ERCOT” — Jim Burke, President and Chief Executive Officer · 2026-08-07. In November: nothing has changed either. Two consecutive quarters of "nothing has changed" while the shares halved is not a reassurance — it is a market that has stopped taking management's load forecast at face value and started marking its own.

The curve is the actual argument

Burke was unusually candid about the mechanism: “we have seen some depression in ERCOT forwards” — James Burke, Chief Executive Officer · 2026-11-06, attributing it to crypto mining facilities going dark to convert into data centers, plus the audit and permitting pause on large loads. And he was explicit that the vaunted $24 does not depend on a rescue: “We are not making any assumptions about power price recoveries” — James Burke, Chief Executive Officer · 2026-11-06. That is either admirable conservatism or a confession that the power price assumption baked into the 2030 target is today's curve, not tomorrow's thesis. The hedge posture reinforces the point. For 2028, “with the 78% hedged for '28, we are much less hedged in ERCOT than the 78%” — James Burke, Chief Executive Officer · 2026-11-06 — deliberately open in the one market where the curve is softest. What has quietly fallen out of the vocabulary is telling: comprehensive hedging was a top-three theme last quarter and vanishes from this quarter's list entirely, replaced by capital allocation. Eight quarters ago this company sold investors on downside protection; today it sells them on share count.

Even after you account for all those commitments that we've announced and we have committed to, we still have roughly $2 billion a year to allocate. And that can go to additional share repurchases or growth projects. For the purposes of this chart, we've made a simplifying assumption that we'll allocate that to share repurchases... we're effectively in this chart, assuming a roughly $3 billion a year of share repurchases.
Kristopher Moldovan, Chief Financial Officer · 2026-11-06

The buyback is now the growth story: 172 million shares retired since November 2021 at an average near $39, a fresh authorization taking the program to about $2.5 billion to be used by no later than 2028, and a stated mid-teens levered return threshold that the company's own paper now clears. It is a strange place for a utility-scale generator to find itself — buying back stock because the market won't pay for growth projects it has already announced.

Beaver Valley, still

Here is the quietest tension on the call. Stacey Dore, on nuclear contracting: “We still have not contracted Beaver Valley. We still have not contracted the second part of Comanche Peak” — Stacey Dore, Chief Commercial Officer · 2026-11-06. Back in February, the same executive had promised that regulatory clarity “will facilitate getting a deal done at Beaver Valley” — Stacey Dore, President and Chief Executive Officer · 2026-02-26. The Beaver Valley keyword was the single highest-ranked theme in 2025Q1 and second-highest in 2026Q1; it does not appear anywhere in this quarter's list. The megawatts are still uncontracted. So are the nuclear uprates beyond the 433 MW tied to the DOE loan. What did close: Cogentrix on October 30, roughly 5,500 MW for about $4 billion net, including 5 million shares struck at an agreed $185 — a valuation the current share price is nowhere near — for approximately $729 per kilowatt. Add Energy Harbor and Lotus and Vistra has closed over 12 gigawatts at under 7x forward EBITDA, with Rumford and Casco Bay already marked for sale. The portfolio is now roughly 50,000 MW, about half of it modern combined cycle. Balance-sheet room is the ballast: $19.6B of non-current debt, up 24% year over year against pro forma net debt to EBITDA of about 2.8x and investment-grade ratings from two of three agencies, with BBB flat as the stated target. Interest coverage of 1.8x is the number to watch if the curve stays depressed.

What the tape is really saying

Vistra is not alone in this de-rating. The broad AI data center complex is among the market's heaviest thirty-day decliners, with the tickers skewing 58 negative to 9 positive — co-packaged optics, high-bandwidth memory, data center capacity all falling together. Talen's call the same week spent its time on large load interconnections and capacity market design, the identical set of unresolved questions Vistra faces on bilateral contracts and PJM's pushed-out backstop auction. The investor's real question, then, is not whether Vistra can generate cash — the record quarter answers that — but whether the large load queue in Texas actually energizes. Management's answer is that raising the bar thins the queue and rewards the projects that clear it, and that its Amazon and Meta contracts clear any bar a legislature could set. The market's answer, so far, is a 43% drawdown. Between those two views sits roughly $2 billion a year of discretionary capital and a CEO who keeps saying the fundamentals didn't change. One of them is early.