The Data Center Boom Is Not Printing Money for Vistra — And the Stock Knows It

I always keep an eye out for the false narratives that sweep through sectors before the structural data catches up. The one circling Vistra right now — that this power company is a guaranteed big winner from the AI data center boom — sounds great in a headline. Today, after Vistra reported its second-quarter earnings, the market has already started dismantling it.

Vistra (NYSE: VST) reported GAAP earnings of $0.91 per share on Thursday, missing the consensus estimate of $2.21 by nearly 60%. Revenue came in at $4.02 billion, well below the $4.93 billion analysts expected. The company's net income for the quarter was $305 million, dragged down by a $472 million unrealized mark-to-market loss on commodity hedges. That hedge loss won't crystallize in cash for years, which is why management rightly highlighted the underlying strength: adjusted EBITDA (earnings before interest, taxes, depreciation, and amortization, a rough proxy for cash earnings) jumped over 30% year over year to $1.77 billion.

That underlying EBITDA growth is real. But it doesn't change what the balance sheet and cash flow statements show, and those are the numbers that survive when narratives shift.

The free cash flow problem

Vistra generated $614 million in free cash flow over the trailing twelve months — down 50.6% year over year. That's the cash left after the company funded $4.06 billion in capital expenditures. Those capex outlays are necessary to build generation capacity, but they eat the cash that would otherwise service debt or grow dividends.

On the other side of the balance sheet sits $35.7 billion in total debt. With equity of just $5.6 billion, the debt-to-equity ratio stands at 341.6%. That is an exceptionally leveraged profile for a company whose free cash flow just halved. A rising interest-rate environment or a sustained dip in wholesale power prices would squeeze that balance sheet hard. The current ratio — a measure of short-term liquidity — sits at 89.6%, below 1.0, meaning Vistra's current assets don't fully cover its current liabilities. Available liquidity as of June 30 was $6.3 billion, which provides a buffer, but the leverage itself is the structural risk.

The dividend that isn't

For an investor buying Vistra as a data center beneficiary, the dividend is the part of the thesis that doesn't add up. The trailing twelve-month yield is 0.64%. That's barely above a savings account. By comparison, Duke Energy yields 3.41%, Southern Company yields 3.10%, and Dominion Energy yields 3.90%. Those are regulated utilities with far more predictable cash flows and fraction of Vistra's leverage.

Vistra's payout ratio — the share of earnings paid out as dividends — is a modest 15%, which gives it room to grow the dividend. But growing from 0.64% takes time, and I trust dividend growth over buybacks as a commitment to shareholders. Vistra has instead been aggressive on buybacks, repurchasing roughly $6.5 billion of stock since November 2021, reducing outstanding shares by about 30%. That's a different capital allocation philosophy, one that benefits remaining shareholders if the stock price holds — but one that offers no income cushion if it doesn't.

The data center pipeline vs. reality

Here's where the false narrative gets tested against the structural data. Vistra has secured impressive long-term contracts on paper. In January 2026, it announced a 20-year power purchase agreement with Meta for over 2,600 megawatts of zero-carbon energy from three nuclear plants. A similar 20-year agreement with Amazon Web Services supports relicensing of Vistra's nuclear facilities. By the end of 2025, Vistra had secured nearly 3.8 gigawatts of 20-year nuclear PPAs with investment-grade buyers. In June, it was named preferred power provider for a $10 billion KKR-backed AI data center venture called Helix Digital Infrastructure, alongside the Kuwait Investment Authority and NVIDIA.

The contracts look formidable. The problem is timing and execution risk. The Meta deal begins delivering power in late 2026, with full capacity not online until 2034. The AWS deal is structured around relicensing, which is a multi-year regulatory process. The KKR-Helix venture is a "preferred provider" relationship, not a closed contract. None of these are printing revenue today.

And then there's the broader demand picture. Energy intelligence firm Currence estimates that 30% to 50% of the large-scale data center capacity expected to come online in 2026 will likely be delayed, due to grid constraints, construction bottlenecks, and permitting delays. Of approximately 16 gigawatts scheduled to begin operations in 2026, only about 5 gigawatts is currently under construction. Roughly 11 gigawatts remains announced with no visible construction progress, despite typical build timelines of 12 to 18 months.

This matters for Vistra because its 2027 earnings guidance is already trending toward the lower end of its $7.4 billion to $7.8 billion opportunity range. Management told investors on today's call that ERCOT (the Texas grid) forward curves are meaningfully lower than what was baked into the original framework. Wholesale electricity prices in Texas averaged $30 per megawatt-hour in Q2 — a level management explicitly said is "not going to get new stuff built." The growth case for 2027 is becoming dependent on the Cogentrix acquisition and Meta PPAs, which management estimates would add roughly $700 million but are excluded from the standalone outlook.

Constellation Energy: same narrative, different results

If Vistra is supposed to be the data center power winner, its closest peer Constellation Energy (NASDAQ: CEG) provides a useful comparison. Constellation also owns nuclear generation and has been hyped as an AI infrastructure play. On August 6, one day before Vistra's report, Constellation posted adjusted EPS of $2.55 versus estimates of $2.29 — a clear beat — and raised its full-year guidance to $11.50–$12.50. The stock popped 2.9%.

But even here, the dividend tells the real story. Constellation yields 0.58%, even lower than Vistra's 0.64%. Its debt-to-equity ratio of 76.4% is far more manageable, and free cash flow of $309 million doubled year over year. The balance sheet is cleaner. But neither stock is an income play, and both are priced for multi-year execution of contracts that haven't hit the cash flow statement yet.

What the stock price has already said

Vistra's shares closed near $142 on Thursday. The stock is down roughly 30% from its 52-week high of $219.82, down 11.7% year to date, and down 17% over the last four months. The market has done what false narratives eventually provoke: it has re-priced the gap between the story and the earnings. Vistra trades at 23.4 times trailing earnings and 11.7 times EV/EBITDA. That's not cheap for a company with a halving FCF, a 0.64% yield, and 342% debt-to-equity. The stock still reflects significant optimism about future data center revenue that is years from materializing.

The conclusion

The AI data center boom is real. The electricity demand it's creating is real. Vistra has real contracts with real hyperscalers. But the false narrative here is that these contracts translate into today's financial results or justify today's leverage. They don't.

The nuclear PPAs with Meta and AWS are structural assets that will matter in 2029, 2032, and 2034 — not in the next earnings quarter. The data center buildout is being delayed at a pace that compresses near-term power price upside. Vistra's debt load is high, its free cash flow is contracting, and its dividend is too thin to compensate for the execution risk. That being the case, I rate Vistra a Sell at current levels. The stock has fallen enough to arrest the most reckless positioning, but it hasn't fallen enough to account for the structural gap between the data center narrative and the cash flow reality. For income investors, regulated utilities like Duke Energy and Southern Company offer yields five times higher with cleaner balance sheets. For growth-oriented investors who believe in the data center thesis, Vistra's risk-adjusted return at 340% debt-to-equity and a sub-1% yield doesn't make the cut.