To The Moon

Mathematical Money | September 29, 2026

$VST Here's a company whose adjusted EBITDA grew 31% year on year last quarter, trading two and a half percent above its fifty-two week low.

I doubled my position in it yesterday. This is the reasoning, including the part that argues against me.

No price target, as usual.

The thing almost nobody accounts for

Vistra has hedged roughly 100% of its expected generation for 2026. About 84% for 2027. About 58% for 2028.

Read that again, because it decides everything else.

Hedging means selling your output forward at a fixed price. It removes the risk of power prices falling. It also removes the benefit of power prices rising.

So the entire "AI data centres are going to make electricity scarce and expensive" thesis — the reason anyone looks at this sector at all — cannot show up in Vistra's 2026 numbers. That year is already sold. It can barely show up in 2027, where only about 16% of generation is open.

It shows up properly in 2028, where roughly 42% is unhedged.

People buy this stock on an AI power story and then get impatient when the earnings don't reflect it. The earnings can't reflect it yet. That's not a disappointment, it's arithmetic that was published in February.

The guidance that leaves out the story

2026 adjusted EBITDA guidance is $6.8 to $7.6 billion, reaffirmed. Free cash flow before growth, $3.925 to $4.725 billion.

For 2027 the company has pointed to an opportunity range of $7.4 to $7.8 billion — and that figure excludes both the Cogentrix acquisition and the Meta PPAs.

That's the sentence I'd underline. The forward number everyone is anchoring to specifically does not include the two assets that constitute the growth case. Whatever those are worth, they sit on top.

I'm not going to pretend I can size that precisely. I can't, and anyone giving you a confident figure is guessing. But "guidance excludes the growth" is a materially different setup from "guidance assumes the growth," and most coverage doesn't make the distinction.

What's actually contracted

Not projected. Signed.

Meta. Twenty-year power purchase agreements for 2,609 MW of carbon-free power from Vistra's PJM nuclear plants — 1,268 MW from Perry, 908 MW from Davis-Besse, plus uprate capacity. Delivery begins on a portion in late 2026, with full delivery by end of 2027.

Twenty years. Not a letter of intent, not a framework — a contracted offtake with a hyperscaler, on nuclear baseload that already exists and already runs.

Taken together with the Amazon arrangement, something close to 3,809 MW of nuclear is under twenty-year contract to those two customers.

Cogentrix. Roughly $2 billion for seven natural gas plants totalling 5,500 MW. FERC approval came through in Q2. Gas is dispatchable — it can ramp up and down — which is the complement to nuclear baseload that runs flat out. Data centres need both: the steady floor and the thing that covers the spikes.

Fleet availability has been running above 97%.

So why is it near its lows?

This is where I have to be honest rather than promotional.

First, it isn't just Vistra. Constellation, the closest comparable, is down about 35.5% from its own high. Vistra is down 36.7%. When two names in the same trade fall by almost exactly the same amount, that's the market repricing the sector, not passing judgement on one company's execution.

Second — and I think this is the real answer — utilities are duration. They're capital-intensive, long-lived, and valued off cash flows stretching decades out. When the long end of the curve went to a nineteen-year high, everything valued that way got marked down. It has very little to do with how much electricity Meta needs in 2028.

That's an uncomfortable thing to say when you own it, because it means the position is partly a bet on rates, and I didn't buy it as a rates bet.

What the chart says

Vistra closed Monday at $138.02. Below its 20-day average. Below its 50-day, 100-day and 200-day. All four, and the 200-day sits 11% above the price.

It's 36.7% below the high of $217.92 set a year ago, and it's sitting at roughly 4% of its fifty-two week range — meaning almost the entire last year of trading happened above where it is now. The low was $134.71 in May and it's about two and a half percent above that.

Sixty-day annualised volatility is roughly 40%.

There's nothing in that chart to like. I want to say that plainly rather than bury it, because the fundamental case and the price action are pointing in opposite directions and only one of them can be right.

How I structured it, and why the date matters

January 2028 calls, deep in the money. I bought two more of them yesterday at $38.80 and wrote four short October calls against the position the same afternoon.

The expiry is not arbitrary. By January 2028 the Meta PPAs are fully delivering, Cogentrix has been owned for over a year, and roughly 42% of generation is unhedged rather than 0%. That is the first point at which the thesis can actually appear in reported numbers rather than in a press release.

Buying a shorter-dated option here would be betting that the market re-rates the story before the story can show up in the accounts. That's a bet on sentiment, and I'm not good enough at sentiment to size it.

Writing short calls against the position is what stops it being dead money while I wait. The long-dated call sits there for sixteen months either way — it may as well collect rent.

What would change my mind

Not the price. Not another quarter of flat earnings, because flat earnings are what a fully hedged year produces by design.

The hedge disclosures. Vistra publishes the percentage hedged for each forward year. If the 2028 figure starts climbing sharply — if they sell the open position forward at prices that don't reflect scarcity — then management is telling you they don't believe in the tightness either, and the whole case dissolves quietly without anyone announcing anything.

The second thing is the delivery schedule. Meta was meant to begin on a portion in late 2026. That's now. If that slips, the 2027 ramp slips with it.

Two things I'd genuinely like other views on.

Does anyone else look at the hedge ladder before buying a power producer? It took me embarrassingly long to realise that a fully hedged year means the commodity thesis is irrelevant to that year's earnings, and I'd have saved myself some impatience by checking it first.

And the harder one — if a sector falls 36% because the discount rate moved rather than because the business deteriorated, is that an opportunity or a warning? I've taken the first view. The chart has taken the second, and it has been right for twelve months.

Stop guessing. Start calculating.

Live to fight another day. 🤙