The chip stocks have been getting all the attention. Meanwhile, one of the most straightforward AI infrastructure plays in the market is sitting near its 52-week low, mostly ignored.
That stock is Vistra Corp. (VST).
Here is the thing about power stocks and AI. Most investors went straight to the nuclear pure-plays when the data center electricity story broke. Constellation Energy got the headlines. Talen got the Amazon deal. But a quieter thesis has been building around Vistra, and it may be more interesting than what is already priced into its better-known peers.
What the Q1 Numbers Said
Vistra reported Q1 2026 GAAP net income of $1.029 billion and Ongoing Operations Adjusted EBITDA of $1.494 billion for the quarter. Management reaffirmed full-year 2026 Ongoing Operations Adjusted EBITDA guidance of $6.8 billion to $7.6 billion and Ongoing Operations Adjusted free cash flow before growth of $3.925 billion to $4.725 billion.
Now here is what makes that guidance range actually interesting: it excludes any contribution from the pending Cogentrix acquisition, which Vistra has said is expected to close in the second half of 2026. And it excludes the 20-year power purchase agreements announced with Meta tied to more than 2,600 megawatts of zero-carbon energy from three Vistra nuclear plants in PJM (Davis-Besse, Perry, and Beaver Valley).
Two potential EBITDA contributors. Neither one is in the guidance range yet.
Why This Is an AI Trade
Hyperscalers have already signed long-duration power purchase agreements with reactor operators. That is not speculative. It is contracted revenue.
What is starting to matter more is something the market has been slow to absorb. AI infrastructure increasingly depends on firm, dispatchable, grid-connected power capable of operating continuously at hyperscale loads. Intermittent renewables cannot do that job around the clock.
Vistra has a diversified generation fleet totaling approximately 44,000 megawatts of capacity.
Slight tangent, but it matters: the semiconductor selloff that has dominated July headlines pulled investors away from anything AI-adjacent that is not a chip stock. When a sector gets punished hard, stocks sharing the same broad theme but with completely different fundamentals often get caught in the same rotation out. That may be part of what is happening with VST right now.
The Valuation Picture
Analyst-count and consensus targets vary by data provider and change frequently. As of mid-July, published consensus estimates generally clustered in the low-$220s to around $230 per share, versus a mid-$150s share price in that same window.
Fitch upgraded Vistra to investment-grade BBB- in March 2026. That upgrade reduces borrowing costs and expands the institutional buyer pool. Those effects take time to show up in price action.
Short-term price performance depends on the exact measurement window and the specific quote used, so treat month-to-date and year-to-date percentages as approximate. The bigger point is that the stock has pulled back despite a strong first quarter and reaffirmed full-year guidance.
What August 7 Brings
Vistra has announced it plans to report second quarter 2026 results on Friday, August 7, 2026. Investors will be watching summer power demand trends in Texas and the mid-Atlantic closely. More importantly, any new commercial agreements with data center customers announced between now and then could move the stock before the report even lands.
The Risk Side
Energy price volatility in ERCOT and PJM can swing quarterly results meaningfully. The Cogentrix acquisition still needs to close. Regulatory or policy changes affecting nuclear-related credits would directly impact the fleet’s earnings trajectory. And valuation multiples move with both price and forward earnings expectations, so any single P/E snapshot should be treated as time-specific rather than definitive.
Those risks are real and visible. But a Meta nuclear PPA outside the guidance range, an August earnings catalyst, an investment-grade credit upgrade, and a stock near a 52-week low is an unusual combination to find in the same name at the same time.
The AI power trade did not end when chip stocks sold off. It just moved somewhere most investors stopped looking.
