Vistra's Revenue "Miss" Is a Red Herring. The Real Story Is in the EBITDA.

The market took Vistra's Q2 revenue decline as evidence the AI-power boom is sputtering. That is a false narrative built on the wrong base year.

Vistra Corp. (NYSE: VST) reported Q2 2026 operating revenue of $4.02 billion, down roughly 5.5% from $4.25 billion a year earlier. The stock slipped about 2% on the news and is down 6.6% over the past week. The consensus narrative, already forming in the headlines, is that the top-line reversal signals a crack in the data-center demand story. The stock has already shed roughly 37% from its 52-week high of $219.82, dragged down by both this earnings print and a broader selloff across AI-power and data-center names as investors grew wary of overbuilding risk.

However, the revenue comparison is a misleading one. Q2 2025 was a price anomaly, driven by an abnormally extreme summer heat wave in ERCOT and PJM markets that sent wholesale power prices to rare peaks. Comparing 2026 to that spike is like measuring growth against a ceiling and calling anything below it a decline.

That being the case, the actual operating story is the opposite of what the headline suggests.

EBITDA Tells the Structural Truth

Adjusted EBITDA — earnings before interest, taxes, depreciation, and amortization, the closest proxy to the cash a power generator actually controls — grew over 30% year over year, rising from $1.35 billion in Q2 2025 to $1.77 billion in Q2 2026. Year-to-date adjusted EBITDA for the first half of 2026 stands at $3.26 billion, up 26% from $2.59 billion in the same period last year.

That growth came from volume expansion, stronger capacity revenue, and contributions from the Lotus Power acquisition, not from a one-time price spike. EBITDA margin sits at 33%, and operating margin is 18.1% — both structurally higher than historical norms for the company.

The $472 million unrealized loss from commodity hedges that hit GAAP net income in Q2 is a non-cash, mark-to-market accounting item that will settle in future years. It does not reflect operational performance. In Q1, Vistra actually recorded a $723 million unrealized hedge gain on the same mechanics, which inflated first-quarter GAAP net income to $1.03 billion. The swing between quarters is volatility in the hedging book, not volatility in the business.

The AI Power Buildout Is Real

Vistra reaffirmed its 2026 adjusted EBITDA guidance of $6.8 to $7.6 billion. More importantly, the company provided a 2027 adjusted EBITDA midpoint of $7.4 to $7.8 billion, excluding any contribution from the pending Cogentrix acquisition or the recently announced Meta power purchase agreements. Those additions would sit on top of that range.

The company is committing up to $1 billion to its Helix Platform — a dedicated data center infrastructure buildout — and announced a partnership involving KKR, KIA, and NVIDIA. These are not speculative moves. They are structural investments in the one secular demand trend the grid sector faces: AI-driven electricity consumption.

Despite the broader market's nervousness about data center overbuilding, the physical constraint is not power demand. It is grid capacity and interconnection timelines. Morgan Stanley's mid-2026 research still points to power constraints emerging by 2027–2028 as the leading commercial bottleneck for AI expansion. Vistra's hedge coverage remains near 100% for 2026 and 94% for 2027, meaning the company has locked in pricing for the near term while building the long-duration contracts that will define its revenue base for the next decade.

What Actually Deserves Scrutiny

The revenue miss is a distraction. The real structural question is balance sheet leverage and free cash flow conversion.

Vistra's total debt stands at $35.7 billion against $5.6 billion in equity, giving a debt-to-equity ratio of 342%. That is steep. Net debt — total debt less $634 million in cash — is $18.5 billion. The company expects to convert over 60% of adjusted EBITDA into adjusted free cash flow before growth capex over the medium term. At current run rates, that would produce roughly $4.1 to $4.6 billion of annual FCF before growth, against guidance.

But the current trailing twelve-month free cash flow figure of $614 million, down 50.6% year over year, reflects the heavy capex cycle the company is in: $4.06 billion in capital expenditures over the past four quarters. This is not deterioration. It is the buildout. Those are the dollars going into Helix, new fleet capacity, and the AI power infrastructure that the 2027 EBITDA guidance is predicated on.

The dividend, at 0.66% yield with a 15% payout ratio, is small. It is growing — six consecutive years of increases, with the latest quarterly raise to $0.23 per share — but VST is not an income stock. It is a growth utility that prioritizes reinvestment and buybacks over shareholder yield. That capital allocation choice matters for EPS leverage but does little for income-focused investors.

Peer Context

Against regulated and hybrid utility peers, Vistra trades at a lower EV/EBITDA multiple of 11.5x versus Ameren at 13.2x, Duke Energy at 14.1x, and NextEra Energy at 18.0x. The discount reflects Vistra's merchant price exposure and higher leverage, but also prices in a growth option that those regulated peers do not have. Vistra's TTM P/E of 22.9x sits above the peer group, but that multiple is distorted by the non-recurring hedge swings that compress GAAP earnings in any given quarter.

None of those peers carry Vistra's direct AI-power optionality or its data-center buildout pipeline. The comparison is imperfect, and in my opinion, it is not fully fair.

The Verdict

The market's focus on a revenue decline measured against an anomalous Q2 2025 is an irrational overreaction that confuses price normalization with structural weakness. The real thesis — EBITDA expansion driven by fleet growth, capacity revenue, and AI power demand — remains intact and is reaffirmed in guidance.

The actual risks are leverage and the execution of the capex buildout. If the grid interconnects materialize, the AI load comes through, and EBITDA hits the $7.4–$7.8 billion 2027 midpoint, the current enterprise value of $65 billion implies a forward EV/EBITDA of roughly 8.4–8.9x. That is cheap for a company with secured offtake and a growing platform.

I rate Vistra a Buy, viewing the sell-off as an irrational overreaction to a meaningless top-line comparison. The stock is an overweight AI-power infrastructure play for investors who can tolerate merchant price volatility and balance-sheet leverage in exchange for structural EBITDA growth. For income-oriented portfolios, the sub-0.7% yield is a nonstarter; for growth-oriented portfolios, the mispriced 2027 optionality is compelling.