Updated 21 August 2026
Vistra (NYSE: VST) traded at $140.10 in the pre-market on 21 August 2026, up 0.83%, after closing at $138.94 on 20 August, down 2.63%, per stockanalysis.com. Market capitalisation is about $46.63 billion and the one-year range runs $132.66 to $219.82.
Verdict: the operating business is compounding and the share price is not. Adjusted EBITDA rose more than 30% year on year in Q2 while the stock fell 28% over twelve months. Street range: bull $313 (+123%), base $219.72 (+57%), bear $106 (-24%). The gap between price and consensus is the story, and it cuts both ways.
Vistra is the awkward stock in the AI-power trade. Everything the bull thesis asked for has arrived – hyperscaler power purchase agreements, a record EBITDA print, a data-centre joint venture with NVIDIA in it – and the shares have gone down anyway. VST changed hands at $140.10 in the pre-market on 21 August 2026, up 0.83% after a 2.63% fall the previous session, and sits roughly 36% below the $219.82 high it set within the past twelve months, per stockanalysis.com. Over those twelve months the stock is down 28.20%.
That is the number to sit with, because the comparison inside its own sector is unflattering in an instructive way. Cameco, the fuel end of the same AI-power complex, is up about 26% over the same twelve months. Vistra, the generation end, is down 28%. Same narrative, opposite tape. Whatever is happening to VST is not a referendum on data-centre electricity demand, because the demand forecasts kept going up the whole time the stock went down.
Key facts
- $140.10 – VST pre-market price, 21 August 2026, 9:11am ET, +0.83%; previous close $138.94, -2.63% – stockanalysis.com
- -28.20% – twelve-month share price change; one-year range $132.66 to $219.82, so the stock sits about 5.6% above its own annual low – stockanalysis.com
- $1,767m – Q2 2026 adjusted EBITDA, up more than 30% from $1.35bn in Q2 2025, on a quarter that missed on revenue – Vistra Q2 2026 results, 7 August 2026
- $6.8-7.6bn – reaffirmed 2026 adjusted EBITDA guidance, with management stating it is comfortable delivering at or above the midpoint; adjusted free cash flow before growth guided to $3.925-4.725bn – Vistra Q2 2026 guidance
- $7.4-7.8bn – the 2027 midpoint opportunity, which excludes roughly $700m from the pending Cogentrix and Meta nuclear agreements – Vistra Q2 2026 call
- ~3,800 MW – nuclear power contracted to Amazon Web Services from Comanche Peak, alongside Meta agreements at PJM nuclear sites – company disclosure
- 13.57x – forward price-to-earnings ratio, against 23.68x trailing and an EV/EBITDA of 10.04x – stockanalysis.com
- $46.63bn – market capitalisation; 335.64m shares outstanding, 85.58% institutionally held, short interest 2.59% – stockanalysis.com
The operating result and the share price have decoupled
Start with what the business actually did. In the second quarter of 2026 Vistra reported adjusted EBITDA of $1,767 million against $1.35 billion a year earlier, an increase of more than 30%. Revenue came in light, which is what most of the same-day coverage led with, but for an independent power producer revenue is the least informative line in the accounts: it moves with wholesale prices and hedging timing in ways that tell you very little about earnings power. EBITDA and free cash flow are the numbers that pay for buybacks, and both went the right way.
Management then reaffirmed 2026 adjusted EBITDA guidance of $6.8 to $7.6 billion and said it was comfortable delivering at or above the midpoint, with adjusted free cash flow before growth of $3.925 to $4.725 billion. For a company with a $46.63 billion market capitalisation, the midpoint of that free cash flow range is a yield most utilities cannot approach. Trailing free cash flow of $2.26 billion works out to a 4.84% yield on the current price.
So the decoupling is real and it is measurable. Over twelve months in which EBITDA grew 30% in the most recent quarter, the equity fell 28%. Multiple compression of that magnitude is not usually a verdict on this year. It is a verdict on what the market thinks the next few years look like.
What actually went wrong: ERCOT prices, not data-centre demand
The most common misreading of VST’s decline is that the AI-power story is deflating. The forecasts say otherwise. PJM utilities project roughly 55 GW of new large load, predominantly hyperscale data centres, by 2030, and something closer to 100 GW by 2037 – against a planning reserve margin already thinning from 18.9% in the 2026/27 delivery year. Nobody serious is forecasting less electricity demand.
The pressure is coming from the other side of Vistra’s book. On the Q2 call management flagged that ERCOT forward prices have softened, offset by strength in PJM and by the hedging programme. Vistra is far more exposed to Texas than to any other market, so a softer ERCOT forward curve compresses the merchant margin that the equity story was capitalising at a high multiple in 2025. That is a genuine, quantifiable negative, and it is the honest core of the bear case. It is also cyclical rather than structural: forward curves move.
The second pressure is contract conversion timing. The AWS agreement at Comanche Peak, roughly 3,800 MW of nuclear, and the Meta agreements across PJM nuclear sites are signed or advancing, but the roughly $700 million that Cogentrix and the Meta nuclear PPAs could add is explicitly not in the 2027 guidance midpoint of $7.4 to $7.8 billion. The market is being asked to underwrite earnings the company has not yet guided to. In a risk-off tape for power names, it declines to.
The Helix joint venture is the option nobody is paying for
The item from the Q2 disclosure that received the least attention may matter most. Vistra committed up to $1 billion to Helix Digital Infrastructure, a data-centre partnership alongside KKR, NVIDIA and KIA, and will serve as preferred power partner to it.
That is a different animal from a power purchase agreement. A PPA sells electrons at a contracted price and caps Vistra’s participation in the economics of the customer’s business. An equity stake in the data-centre platform, plus preferred supplier status, converts the generator from a commodity vendor into a participant in the compute build-out itself. Whether it earns its cost of capital is unknowable today, and prudent investors should treat the $1 billion as capital at risk rather than value created. But it is an option with real convexity, and at 13.57x forward earnings the market is assigning it approximately nothing.
Scenarios: bull $313, base $219.72, bear $106
Against the $140.10 pre-market price, the published street range is unusually wide, which is itself the most honest description of the setup. Nineteen analysts cover the stock with a Strong Buy consensus, and their targets span $106 to $313 – a spread of more than 2.9 times from low to high.
| Case | 12-month level | vs $140.10 | Anchor |
|---|---|---|---|
| Bear | $106 | -24.3% | The lowest published target of the 19 analysts covering VST (stockanalysis.com). Assumes ERCOT forwards stay soft, Cogentrix and the Meta nuclear PPAs slip, and the merchant multiple compresses toward regulated-utility levels. Note this sits well below the $132.66 one-year low, so it requires a genuine break of the current range. |
| Base | $219.72 | +56.8% | The consensus average target across 19 analysts (stockanalysis.com). It also lands almost exactly on the $219.82 one-year high, so consensus is effectively forecasting a full round trip. Requires 2026 EBITDA at or above the guided midpoint and the pending PPAs converting into 2027 guidance. |
| Bull | $313 | +123.4% | The highest published street target (stockanalysis.com). Requires ERCOT forwards to recover, the full ~$700m of pending contract EBITDA to land, and the Helix stake to be valued as a growth asset rather than capital expenditure. |
A word of caution about that base case, because a consensus target 57% above the traded price is not a normal state of affairs. It usually resolves one of two ways: the targets come down, or the price goes up. Recent revisions have leaned toward the second. Morgan Stanley’s David Arcaro raised his target to $227 from $212 on 21 August while keeping an Overweight rating; Exane BNP Paribas’ Moses Sutton published $255 on 19 August, TD Cowen’s Shelby Tucker $221 on the same day, and DBS’ Pei Hwa Ho $216 on 18 August. Bernstein sits lowest of the recent cluster at $181, still 29% above spot. Four separate revisions in four sessions, all above the market price, is a sell side that has not capitulated.
The counterpoint a buyer should hold onto: the same sell side was equally constructive at $219 before the stock lost 28%, and the $106 low target exists for a reason. Analyst dispersion this wide means the honest answer is that the outcome depends on the ERCOT forward curve, which nobody in the coverage list controls.
Quick Take
The bull case in one line: a business growing EBITDA 30% year on year, trading at 13.57x forward earnings, with roughly $700m of contracted upside not yet in guidance and an NVIDIA-adjacent data-centre stake valued at zero.
The bear case in one line: soft ERCOT forward prices hit the merchant margin that justified the old multiple, the pending PPAs are promises rather than guidance, and the stock is only 5.6% above its annual low for a reason.
What decides it: the ERCOT forward curve and whether Cogentrix and the Meta nuclear agreements convert into the 2027 guidance number. Watch the next guidance update, not the next demand forecast.
How Vistra sits against the rest of the power complex
It is worth placing VST beside its peers rather than reading it alone, because the AI-power trade has stopped moving as one block. Cameco is up about 26% over twelve months on the fuel side. Constellation Energy and the wider generation group have gone the other way, with CEG about a third below its own high. Smaller reactor names such as NuScale trade on a pipeline that converts slowly, and distributed-generation names such as Bloom Energy have followed a separate path again.
The dispersion tells you the market is no longer buying “AI needs power” as a single trade. It is discriminating by fuel, by market, and by contract structure. Vistra’s discount is a Texas merchant-pricing discount, not an AI-demand discount, and that distinction is the one that determines whether the current price is an opportunity or a warning.
Frequently asked questions
Why is Vistra stock down if data-centre power demand is rising?
Because the two are less connected than the narrative implies. Vistra’s earnings are driven substantially by merchant power prices in ERCOT, and management flagged softer ERCOT forward prices on the Q2 2026 call. Demand forecasts for PJM rose over the same period. The share price is tracking the price of the electricity Vistra sells, not the quantity the market expects to need.
What is Vistra’s 2026 guidance?
Adjusted EBITDA of $6.8 to $7.6 billion, reaffirmed at the Q2 2026 results on 7 August, with management stating it is comfortable delivering at or above the midpoint. Adjusted free cash flow before growth is guided to $3.925 to $4.725 billion.
What is the analyst price target for VST?
The consensus average is $219.72 across 19 analysts, with a Strong Buy consensus rating, per stockanalysis.com as at 21 August 2026. The range runs from $106 at the low to $313 at the high. The most recent revision is Morgan Stanley’s David Arcaro at $227, raised from $212 on 21 August.
Is Vistra cheap at these levels?
On forward earnings it screens inexpensively for the sector at 13.57x, with an EV/EBITDA of 10.04x and a trailing free cash flow yield of 4.84%. On trailing earnings it is 23.68x. Whether that is cheap depends entirely on whether forward EBITDA holds, which in turn depends on ERCOT forward prices. A low multiple on an earnings number that falls is not a low multiple.
What are the Meta and Amazon agreements worth to Vistra?
The company has contracted roughly 3,800 MW of nuclear power to Amazon Web Services from Comanche Peak, and has agreements with Meta at PJM nuclear sites. On the Q2 call management indicated the Meta nuclear PPAs together with Cogentrix could add approximately $700 million to future guidance. That amount is not included in the 2027 EBITDA midpoint of $7.4 to $7.8 billion.
What is Helix Digital Infrastructure?
A data-centre partnership involving KKR, NVIDIA and KIA, to which Vistra has committed up to $1 billion and in which it will act as preferred power partner. It gives Vistra an equity interest in data-centre infrastructure rather than only a contracted supply relationship. It is early stage and the capital should be treated as at risk.
How far is VST from its highs and lows?
The one-year range is $132.66 to $219.82. At the $140.10 pre-market price on 21 August 2026, the stock is roughly 36% below its annual high and about 5.6% above its annual low.
Price and valuation data in this article was verified on 21 August 2026 from stockanalysis.com; operating figures and guidance are from Vistra’s Q2 2026 results and earnings call of 7 August 2026. Prices move; check a live quote before acting on anything here.
This article is for information only. It is not investment advice, and it is not a recommendation to buy or sell any security. FinanceFeeds does not hold positions in the companies mentioned. Do your own research and consider speaking to a regulated financial adviser before making investment decisions.