Vistra (VST): AI Power Play or Overpriced Utility?
Every AI-power bull owns this one. We ran VST through the three gates anyway — and the answer isn't the slam dunk the narrative suggests.
Why this one's back in the conversation
VST is the stock everyone wants to talk about again. Vistra was just chosen as the preferred power provider for Helix Digital Infrastructure, a new $10 billion KKR-backed venture supporting AI-focused hyperscale data centers, which dropped it back onto the front page this week.
And yet the chart isn't acting like a hype stock. The 30-day return is up 3.79%, but shares are down 6.88% over 90 days and off 10.42% year-to-date. That gap — euphoric narrative, lukewarm tape — is exactly the kind of setup our model is built to adjudicate. So let's run it.
Survival: 93 — the balance sheet finally caught up
This is the highest of the three gates, and the fundamentals back it up. Vistra hit investment-grade ratings at both S&P and Fitch, with Fitch's BBB- upgrade citing an improved business profile, strong credit metrics, supportive capital allocation, and better market fundamentals. For a company that emerged from bankruptcy a decade ago, that's a meaningful rerating of credit risk.
Hedge coverage sits at roughly 98% for 2026, 89% for 2027, and 65% for 2028 — meaning near-term cash flows are largely locked in regardless of where power prices drift. Pair that with about $4.17 billion of liquidity as of March 31, 2026, and the "can this company survive a power-price air pocket?" question gets harder to answer in the negative.
Growth: 63 — real, but not WOW
A 63 here is the most interesting score, because the AI-power crowd will tell you growth deserves a 95. The actual numbers say something more nuanced. Q1 2026 adjusted EBITDA was $1.494 billion, up roughly 20% year-over-year and 85% from Q1 2024 — that's a genuinely strong print, not a rounding error.
The contracted backlog is where the AI thesis lives. Meta locked in two 20-year PPAs covering 2,609 MW of carbon-free generation and capacity from Vistra's nuclear fleet in PJM, and the company announced the 5,500 MW Cogentrix natural gas acquisition and signed long-term PPAs with Meta and AWS. That's the bull case in three sentences.
So why isn't Growth scoring higher? Two reasons. First, the timing — the Meta PPA more than triples the plant's output, with power ramping from late 2027 through 2032. The cash flows are real, but they're years out. Second, current guidance is already chunky: 2026 adjusted EBITDA guidance of $6.8–$7.6 billion and adjusted FCFbG guidance of $3.925–$4.725 billion, with a 2027 midpoint opportunity of $7.4–$7.8 billion, excluding Cogentrix and new PPAs. There's growth, but the step-function from here is mid-single digits on the base business until the new PPAs actually start delivering.
Valuation: 67 — not cheap, not absurd
The question in our headline isn't rhetorical — 15x EBITDA is rich for a power company by historical standards, and that's the bear case in one number. The model gives back a 67, which sits in "fair-to-slightly-rich" territory rather than "sell on sight."
The case for paying up: consensus expects EBITDA to reach $7.34B, the contracted PPA book stretches 20 years, and the buyback program is aggressive — about $6.3 billion of share repurchases since November 2021, reducing shares outstanding to roughly 337 million, with about $1.5 billion of authorization still expected to be completed by year-end 2027. Compounding stories with that kind of capital return discipline rarely trade at "cheap" multiples.
The case against: key risks include hyperscaler CapEx volatility, regulatory intervention, and valuation sensitivity despite the recent pullback from 52-week highs. If AI capex even normalizes — not collapses, just normalizes — a 15x multiple becomes very hard to defend. The thesis is wrong if hyperscaler power demand turns out to be front-loaded rather than structural, or if regulators decide "behind-the-meter" nuclear deals shouldn't have the economics Vistra is currently underwriting.
The bottom line
GOOD, not WOW. Vistra is a Quality Compounder — investment-grade balance sheet, multi-decade contracted demand, real EBITDA growth, and management returning capital. That's a defensible business.
But the AI-power narrative has done most of the heavy lifting on the multiple already. A 63 on Growth and a 67 on Valuation are the model's polite way of saying: the bull case is largely in the price, and the marginal buyer here is paying for execution on PPAs that don't fully ramp until 2027–2032. Worth owning the business? Maybe. Worth chasing it on the next Helix-style headline? That's a different question — and one only you can answer for your own portfolio. This is analysis, not advice.