Vistra (VST) Deep Dive: Nuclear Moats, $20B of Debt, and What It's Actually Worth
The bull case is real. The debt is realer.
Vistra (NYSE: VST) is the largest independent power generator and retail electricity provider in the US. 44GW of capacity (27GW gas, 6.4GW nuclear, 8.7GW coal), 5 million retail customers, ~$4.3 billion in annual free cash flow before growth. The stock went from $40 in early 2024 to $200+ by late 2025 on the AI-data-center-needs-nuclear narrative, then pulled back to $149–160 today.
This post tries to answer one question: what is it actually worth?
The Business
VST both generates and sells electricity — think farmer and grocer in one. The generation side earns the spread between wholesale power prices and fuel costs. The retail side charges households and businesses for electricity, earning the markup between wholesale purchase and retail sale.
The combination matters more than either piece alone. When power prices spike, generation profits surge but retail gets squeezed (buying power is expensive). When prices drop, generation earns less but retail margins expand (procurement is cheap). The two offset each other, making total cash flow materially more stable than a pure merchant generator.
But don’t overrate this hedge. Q1 2026 was a warm winter: Retail EBITDA dropped from $184M to $68M year-over-year. Weather shifts, profits swing.
Where the Money Comes From
Top-line revenue of $14.3 billion from the retail segment looks large but is misleading — most of it is pass-through procurement cost. Retail EBITDA is only ~$1.4–1.6 billion. The real profit engine is generation, contributing ~$4.3 billion in EBITDA.
Breaking down the 2026 guided EBITDA of ~$7.2 billion by source: nuclear baseload contributes roughly $1.8 billion (fuel cost just $5–7/MWh, margins are enormous), gas-fired spark spread earnings around $2.0 billion (98% hedged for 2026), PJM capacity payments about $1.3 billion (auction-locked for 2–3 years, paid regardless of actual generation), retail ~$1.4 billion, and the rest from corporate and other items.
Then the waterfall down to what common shareholders can actually take home: subtract $1.13 billion in interest, $1.54 billion in maintenance capex / nuclear fuel / long-term service agreements, $160 million in preferred dividends, $130 million in stock-based comp, and $100–200 million in Asset Closure cash costs. Clean common owner earnings: roughly $3.9 billion — about 54% of EBITDA. The rest gets eaten by interest (16%), maintenance (21%), and other senior claims (9%).
The Nuclear Contracts
AWS signed a 20-year PPA for 1,200MW from the Comanche Peak nuclear plant in Texas, starting Q4 2027 and ramping to full load by 2032. Meta signed for 2,176MW from the Perry and Davis-Besse plants in Ohio (full load by end of 2027) plus 433MW of uprates across three eastern nuclear stations (2031–2034). At full delivery, these contracts add $9–12 billion in incremental annual cash flow — 20–25% on top of current FCFbG.
But every critical economic term is undisclosed: the price per MWh, escalation mechanisms, take-or-pay provisions, replacement power obligations, and uprate capital costs. The AWS contract contains the phrase “if the Customer utilizes the full capacity” — a conditional, not an unconditional commitment.
The non-disclosure is rational. AWS and Meta don’t want competitors seeing their power cost structure. VST still has ~3.1GW of uncontracted nuclear capacity and doesn’t want to anchor the next buyer’s price expectations. And publishing a high price would invite political heat — “why do tech giants get to lock up scarce nuclear power while household bills rise 30%?”
You know the house is rented. You don’t know if the rent is good or bad.
The 10x Capacity Price
PJM capacity auction prices went from $29/MW-day to $333 — driven by coal retirements, construction bottlenecks (gas turbine order books are full through 2030), AI data center load growth, and electrification. For VST, ~10.5GW of cleared capacity at $333 means roughly $1.3 billion per year in standby payments, collected regardless of how much electricity is actually generated.
$333 is almost certainly not an equilibrium price — at this level, new-build gas plants earn 15–20% returns, which will attract supply. But physical constraints mean that supply response takes 4–5 years minimum. The excess profit window is temporary, but “temporary” can last a long time.
There’s an underappreciated bifurcation here. Gas plants’ excess profits are cyclical — new gas will eventually arrive to compete them away. Nuclear excess profits are structural — you cannot build new nuclear in any foreseeable timeframe (Vogtle took 15 years and $35 billion). Existing nuclear plants are irreplaceable assets. If the market is pricing nuclear and gas in the same “capacity sensitivity” framework, nuclear is likely being systematically undervalued.
Layered Valuation
Applying a single multiple to $4.3 billion of FCFbG is crude. A dollar of 20-year nuclear PPA cash flow has nothing in common with a dollar of unhedged merchant gas exposure. They deserve completely different multiples.
Current operating assets break down as follows. Nuclear at 6.4GW is irreplaceable, with licenses running to the 2040s–2060s — valued at 12x EBITDA, that’s $21.6 billion. The 22GW CCGT gas fleet, benchmarked to the Cogentrix transaction at 7x, comes to $14.0 billion. Lotus (2.6GW gas), peakers, and coal add another $5.1 billion at lower multiples reflecting their age and declining profile. The retail platform at 7x normalized EBITDA is $9.8 billion. Locked-in PJM capacity for the next two years discounts to $2.4 billion. Net of corporate overhead, current operating enterprise value is roughly $51.4 billion.
Signed PPA incremental value, discounted at 10% (well above utility-grade rates, reflecting undisclosed pricing), adds $6.8 billion. Uncertain items — remaining nuclear recontracting, long-term capacity premium, Cogentrix net accretion — contribute another $4.8 billion if you give them partial credit.
Total enterprise value: ~$63 billion.
The Debt Reality
This is where the analysis changed the most. Pulling the Q1 2026 10-Q balance sheet, the actual priority claims are far larger than I initially estimated.
Gross debt — long-term debt including current maturities, accounts receivable financing, and the forward repurchase obligation — totals $20.6 billion. Unrestricted cash is just $634 million (restricted cash of $43 million doesn’t count). Net debt: $19.9 billion. Add preferred stock at $2.5 billion liquidation value across three series (7.0%, 8.0%, and 8.875% coupons), plus roughly $800 million in net ARO exposure after accounting for the $5.0 billion nuclear decommissioning trust that covers most of the nuclear retirement liability. Total priority claims: $23.2 billion.
Out of a $63 billion enterprise, $23.2 billion must be paid to others before common equity sees a dollar. Common equity value: $39.8 billion.
And Cogentrix hasn’t closed yet. That transaction adds ~$2.3 billion in cash consideration plus $1.5 billion in assumed debt plus 5 million shares of dilution. Post-close, net debt could jump to $23 billion or higher.
At this leverage, every 1% move in enterprise value translates to roughly 1.6% move in equity. The lever amplifies everything — upside and downside alike.
What It’s Worth
The midpoint expectation is roughly $128/share ($63 billion enterprise value minus $23.2 billion priority claims, divided by 337 million shares). Current price of $149–160 implies a 16–25% premium to that estimate. Not a bubble, but no margin of safety either.
If you strip out everything uncertain — zero credit for nuclear recontracting, zero for long-term capacity premium, zero for Cogentrix accretion, and haircut the signed PPAs by half — you get enterprise value of ~$54.4 billion. Apply the actual priority claims from the 10-Q at $23.2 billion, and common equity is $31.2 billion, or roughly $95/share. At that price, every future growth driver is free.
VST is not a bad company. The nuclear fleet is genuinely irreplaceable. The capacity scarcity is real. The PPA contracts, if well-priced, could transform the business from a cyclical generator into a contracted infrastructure platform. But $19.9 billion of net debt means common shareholders stand far back in the capital structure. At today’s price, you need nearly every optimistic assumption to work simultaneously to earn a return.
The right move is to wait for a better price.
Disclaimer: This is not investment advice. All valuations are based on public information and a personal analytical framework. Actual outcomes may differ materially.