Vistra Sold Its Nuclear Expansion Before Washington Funded It

When Big Tech sneezes on capital expenditures, merchant power gets pneumonia. On October 22, Meta delivered a quarterly update featuring aggressive infrastructure outlays and negative free cash flow. In response, Vistra slumped nearly 10% on more than double its average trading volume, falling twice as hard as its independent peers while regulated utilities actually climbed. Investors dumped the country’s largest competitive power generator as if it were an unprofitable software subscription, conflating tech balance-sheet anxiety with power-grid fundamentals.

Strip away the trading screens and Vistra’s business is refreshingly mechanical. The company generates electricity across 43,641 megawatts of capacity, spanning six nuclear reactors, 28 combined-cycle natural gas facilities, fast-start peakers, and solar arrays across Texas and the mid-Atlantic. It also sells power straight to residential meters through retail arms like TXU Energy. Unlike a traditional regulated utility that earns an approved return on state-sanctioned construction budgets, Vistra plays the open wholesale market. It profits on the spread between the cash cost of splitting an atom or firing gas and the market clearing price for electrons. Think of Vistra as a commercial landlord who owns the only high-rise with heavy industrial wiring, charging premium rents while everyone else waits years for the local utility to install a substation.

That dynamic makes Vistra the primary toll collector of the artificial intelligence buildout. Frontier models require around-the-clock baseload electricity that solar panels and wind farms cannot reliably deliver alone. Because connecting a brand-new generating station to the grid takes years of interconnection studies and severe hardware backlogs, an operational nuclear plant with an active switchyard is the scarcest asset in technology. That structural bottleneck gives Vistra pricing leverage. In January, Meta signed 20-year power purchase agreements covering 2,609 megawatts across Vistra’s nuclear fleet in Ohio and Pennsylvania, locking in existing capacity alongside future additions.

Vistra sold its scarcest expansion forward before it even secured the money to build it. The Meta transaction includes 433 megawatts of nuclear uprates—complex plant modifications that coax additional thermal energy from an existing reactor core. In early October, the Department of Energy backed that commercial deal with a conditional loan commitment of up to $4.2 billion to modernize those plants and support operating them an extra 20 years. Vistra contracted the future power at a corporate premium, then secured non-dilutive government financing to fund the works.

The bottleneck on these new megawatts is no longer bank credit. It is the docket at the Nuclear Regulatory Commission. Formal license amendment filings run from 2029 to 2032, and the first extra electrons reach the grid around 2031.

Short sellers, however, see an equity whose core business is decelerating under substantial leverage. Exclude the pending Cogentrix gas acquisition and the Meta contract premium, and 2027 adjusted EBITDA expands by only a few percent over 2026. That sits against more than $19 billion in net debt. Servicing that debt is growing more expensive. Vistra recently issued $1.5 billion in junior subordinated notes yielding over 7%. It also tapped $1.0 billion of cash to redeem preferred stock, leaving only about $435 million on the balance sheet.

Meanwhile, power prices in Texas are deflating. Forward prices in ERCOT for 2027 settled near $46 per megawatt-hour in early October, down 22% year over year. While Vistra holds profitable hedges averaging nearly $59, those contracts eventually expire.

Actual data-center power draw is lagging behind paper commitments. ERCOT has approved nearly 9,500 megawatts of large industrial loads, but peak observed use sat below 4,400 megawatts this summer. Texas has since paused new data-centre connection studies. If Texas forward curves fail to rebound or reactor filings slip, the bear thesis holds that Vistra capped its upside just as merchant power peaked.

For the equity to re-rate from here, management must prove its growth is real and decoupled from hyperscaler sentiment. The primary catalyst arrives on November 6 with third-quarter earnings. Executives previously indicated that closing Cogentrix and adding the Meta premium will layer roughly $700 million onto the 2027 EBITDA midpoint. That is a 9% boost, and it needs to appear in the formal guidance ledger, not just an earnings call transcript. That revision would materially compress Vistra's forward multiple on numbers Wall Street can model. Beyond that print, investors need to see ERCOT power forwards establish a floor and confirmation that license extension filings for the Beaver Valley and Davis-Besse reactors remain on schedule.

We are watching the November 6 earnings release to see if management officially moves its 2027 adjusted EBITDA midpoint past $8.1 billion. The stock trades like a high-beta technology proxy, but its long-term revenue is anchored by binding, multi-decade purchase contracts with investment-grade buyers. Vistra does not need Meta's stock to rally to collect its cash.

Disclosure: The Signal holds no position in VST. Positions may change. This is not financial advice.