Vistra is the largest independent electricity producer in the United States, selling power to homes and businesses in Texas and the Northeast while running one of the country's biggest nuclear fleets. It has signed 3,800 megawatts of twenty-year electricity supply agreements with Meta and Amazon, and the stock has fallen 37% from its high while those contracts were being signed.
Catalysts
Contracted revenue starting to flow. Roughly 3,800 megawatts under twenty-year agreements with Meta (over 2,600 MW across three nuclear plants) and Amazon Web Services. Around half of company profits are shifting toward contracted and household customers rather than volatile market sales.
Perry nuclear plant restart. Adds generating capacity without building anything new.
The Cogentrix purchase. Roughly 5,500 megawatts of modern gas-fired plants for approximately $4 billion, expected to complete in the second half of 2026 subject to regulatory approval.
Continued share buybacks. About 30% of shares retired since late 2021, with $1.5 billion of buyback authority remaining and more than $10 billion of cash expected to be generated through 2027.
A partnership with NVIDIA. Vistra has committed up to $1 billion to Helix Digital Infrastructure alongside KKR, NVIDIA and the Kuwait Investment Authority, as founding investor and preferred power supplier.
Credit rating upgrade. Fitch upgraded Vistra's debt to investment grade, the second major agency to do so, which lowers future borrowing costs.
Key risks
Dominant risk: politicians acting on electricity bills. As data centres compete for power, household electricity costs rise, and governments respond. New York has already introduced a one-year ban on new data centres using 50 megawatts or more. Texas and the eastern grid operator are reviewing similar concerns. Anything that encourages more power supply reduces the scarcity that currently makes Vistra's plants valuable.
Texas electricity prices are weak. Softer expected prices led management to say 2027 profits are tracking toward the lower end of their forecast range.
Price protection runs out in 2028. Vistra has locked in prices for roughly 100% of its 2026 output and 94% of 2027, but only 72% of 2028. That gap sits inside a three-to-five year holding period.
Substantial debt. Roughly $23.5 billion net of cash, about $70 per share, or half the share price.
Accounting profits are falling even as cash profits rise. In the June quarter, reported earnings per share fell to $0.77 from $0.82, revenue fell 5.5% to $4.02 billion, and net income fell 7.9% to $258 million, while the cash-based profit measure rose over 30%. Both facts are true and the tension is unresolved.
Portfolio position
1.2% of TFSA+2.68%
Vistra Corp (VST) Investment Evaluation Brief
Last updated: August 25, 2026
Price at writing:$138.02 (August 25, 2026). The stock is volatile: it has traded between $138 and $153 in the past week alone, and has fallen 27.84% over 52 weeks.
Market cap:$46.32B (335.64M shares outstanding)
Sector: Utility / Independent Power Producer
Layer: Power Generation
Conviction: Medium (weighted score 3.93 / 5.0; expected hold 3 to 5 years)
Verdict: Buy under ~$145, then deploy double the cash under ~$112. At $138.02 the first line is live, and the stock sits 37% below its 52-week high of $219.82.
Vistra is the largest independent electricity producer in the United States, selling power to homes and businesses in Texas and the Northeast while running one of the country's biggest nuclear fleets. It has signed 3,800 megawatts of twenty-year electricity supply agreements with Meta and Amazon, and the stock has fallen 37% from its high while those contracts were being signed.
1️⃣ Summary/Snapshot
Vistra owns nuclear plants that already exist and already run, and it has agreed to sell their output to two of the world's largest technology companies for the next twenty years. The plan: buy under ~$145, then deploy double the cash under ~$112.
A simple analogy: most power companies are like farmers selling crops at whatever the market pays on harvest day. Vistra is a farmer who has pre-sold a large share of the harvest at a fixed price for twenty years to two very reliable buyers, and who also owns the grocery store where the rest gets sold.
Contracts, not construction. The AI-related revenue comes from nuclear plants that are already built and running. Nothing new needs to be constructed, no money needs to be raised, and no regulator needs to approve anything. Compare that to competitors promising power from plants that do not yet exist.
A large part of the business does NOT depend on AI. Roughly half of Vistra's operating profit comes from selling electricity to ordinary households and businesses. That revenue continues whether or not data centre construction slows down. It lowers exposure to the single biggest risk shared across every AI-related investment.
Valuation. Worst case $112, base case $151, best case $196, with a probability-weighted fair value of $154 against a price of $138.02 (August 25). Today is a reasonable entry for a first purchase, but it is not a bargain.
It pays a dividend of $0.92 per share per year, a yield of 0.67% at the current price. That is small, because the company returns most of its cash by buying back its own shares instead. It has bought back roughly 30% of all shares outstanding since late 2021, which increases each remaining shareholder's stake.
Who else owns it. Thiel Macro, the investment fund of Peter Thiel (the billionaire investor who co-founded PayPal and Palantir and was Facebook's first outside investor), increased its holding from 208,747 to 372,755 shares in the second quarter of 2026. It is that fund's third-largest disclosed position.
When to sell, in plain terms
If this happens
Do this
Meta or Amazon cancel or substantially rewrite their electricity contracts
Sell the whole position
Texas or the eastern states block new data centres, as New York has just done
Sell half, then reassess
Credit rating agencies downgrade the company's debt to below investment grade
Sell half
Texas electricity prices stay very low into 2028 while the company's price protections run out
Sell half
Texas power prices stay below $40 per megawatt-hour through the autumn results
Stop buying more, but do not sell
At a Glance
Verdict
Buy under ~$145, double under ~$112
Current price
$138.02 (August 25, 2026)
Base-case fair value
$151 (+9.4%)
Probability-weighted fair value
$154 (+11.6%)
Analyst average / median
$225.09 / ~$225 (23 analysts)
Moat
An existing nuclear fleet that cannot be copied
Key catalyst
Contracted revenue ramping, Perry plant restart
Risk to watch
Politicians acting on rising household electricity bills
Dividend
$0.92/yr per share (0.67% yield)
Next earnings
Early November 2026
Valuation Summary
Scenario
Fair Value / Share
vs $138.02 (August 25)
Conservative
$112
-18.9%
Base
$151
+9.4%
Bull
$196
+42.0%
Probability-weighted (20/55/25)
$154
+11.6%
Notable Analyst Price Targets
Firm
Rating
Target
Date / notes
Consensus (23 analysts)
Buy, 83%
$225.09
Revised down 0.46% over three months
Wells Fargo
Overweight
$212
Lowered from $220 after the summer results
Bernstein
Outperform
$187
Initiated, citing diversified generation
Simply Wall St (model)
n/a
$221.57
Consensus fair value estimate
The Street sits far above every scenario here. That gap is a difference in method, not a disagreement about the business: analysts apply more optimistic profit multiples than this brief does. Worth noting the direction of travel is downward, with targets trimmed over the past three months.
Where to Find More Detail
Business explanation, Section 4. Valuation, Section 6. Scoring, Section 7. Purchase schedule and exit rules, Section 8.
2️⃣ Catalysts
Contracted revenue starting to flow. Roughly 3,800 megawatts under twenty-year agreements with Meta (over 2,600 MW across three nuclear plants) and Amazon Web Services. Around half of company profits are shifting toward contracted and household customers rather than volatile market sales.
Perry nuclear plant restart. Adds generating capacity without building anything new.
The Cogentrix purchase. Roughly 5,500 megawatts of modern gas-fired plants for approximately $4 billion, expected to complete in the second half of 2026 subject to regulatory approval.
Continued share buybacks. About 30% of shares retired since late 2021, with $1.5 billion of buyback authority remaining and more than $10 billion of cash expected to be generated through 2027.
A partnership with NVIDIA. Vistra has committed up to $1 billion to Helix Digital Infrastructure alongside KKR, NVIDIA and the Kuwait Investment Authority, as founding investor and preferred power supplier.
Credit rating upgrade. Fitch upgraded Vistra's debt to investment grade, the second major agency to do so, which lowers future borrowing costs.
3️⃣ Key Risks
Dominant risk: politicians acting on electricity bills. As data centres compete for power, household electricity costs rise, and governments respond. New York has already introduced a one-year ban on new data centres using 50 megawatts or more. Texas and the eastern grid operator are reviewing similar concerns. Anything that encourages more power supply reduces the scarcity that currently makes Vistra's plants valuable.
Texas electricity prices are weak. Softer expected prices led management to say 2027 profits are tracking toward the lower end of their forecast range.
Price protection runs out in 2028. Vistra has locked in prices for roughly 100% of its 2026 output and 94% of 2027, but only 72% of 2028. That gap sits inside a three-to-five year holding period.
Substantial debt. Roughly $23.5 billion net of cash, about $70 per share, or half the share price.
Accounting profits are falling even as cash profits rise. In the June quarter, reported earnings per share fell to $0.77 from $0.82, revenue fell 5.5% to $4.02 billion, and net income fell 7.9% to $258 million, while the cash-based profit measure rose over 30%. Both facts are true and the tension is unresolved.
4️⃣ Plain-English Business Explanation
What the company does
Vistra makes electricity and sells it. That is the whole business, and it operates at both ends of the chain.
At one end it owns power plants: nuclear reactors, natural gas plants, coal plants and large batteries, spread across Texas and the eastern United States. These generate electricity that gets sold into wholesale markets, where prices move constantly with weather, demand and fuel costs.
At the other end it owns retail electricity brands that sell directly to households and businesses. When a Texas family chooses an electricity provider, one of the options is a Vistra brand. That retail business buys power and resells it, earning a margin on the difference.
Owning both ends is deliberate. When wholesale prices spike, the plants earn more while the retail business pays more. The two partly cancel out, which makes total profits steadier than either half would be alone.
A note on the Texas market
Most of Vistra's profits come from Texas, which runs its own electricity grid separate from the rest of the United States. The operator is called ERCOT, the Electric Reliability Council of Texas.
Texas is unusual in one important way. In most American electricity markets, generators are paid twice: once for the electricity they actually produce, and again simply for being available when needed. Texas has no second payment. Generators earn only from electricity they actually sell.
The consequence is that Texas profits swing hard with weather. A hot summer means heavy air conditioning use, tight supply, high prices and strong profits. A mild summer means the opposite, with nothing to cushion the fall. This is why "Texas power prices" appear repeatedly as a risk in this brief.
Business segments
Generation. Nuclear, gas, coal and battery plants selling into wholesale markets. June quarter revenue of $994 million, up 68%.
Retail. Electricity sold directly to end customers under consumer brands. June quarter revenue of $773 million.
How it connects to the AI build-out
The narrow connection: Data centres need enormous amounts of electricity, available continuously, day and night. Nuclear plants are the only large-scale source that runs around the clock without emissions. Vistra has agreed to sell that output to Meta and Amazon for twenty years.
The wider connection: Data centre demand lifts baseline electricity use across whole regions. That tightens supply and raises prices for every generator, including those with no data centre contracts at all.
Is the core driver AI-dependent? Partly. The retail business, roughly half of profits, does NOT depend on AI. It serves ordinary customers who would buy electricity regardless. The twenty-year contracts are locked in, so they survive even a sharp slowdown in data centre construction. What is AI-dependent is the upside: the portion of electricity sold at market prices that benefits from AI-driven scarcity.
Who uses it: Texas and northeastern households and businesses buying retail electricity, plus Meta and Amazon Web Services buying nuclear output under long-term contract.
Why the market has not fully priced it
Reported profits look bad. Revenue and net income both fell in the June quarter while cash profits rose 30%. Automated stock screens that rank on reported earnings show deterioration.
The 37% price drop reads as a broken story. In fact the major contracts were signed during the decline.
Regulatory fear is being applied to the whole sector. The New York ban and Texas scrutiny create a broad discount, but Vistra's contracted revenue is largely protected from it.
Market-exposed power companies carry a permanent discount for commodity risk, and investors have not yet separated Vistra's contracted revenue from its market-exposed revenue.
Key business metrics
Metric
Value
Contracted capacity under long-term agreement
3,800 MW, 20-year terms
Price protection in place
~100% for 2026, 94% for 2027, 72% for 2028
Share count reduced since 2021
~30%
Return on equity
42.96%
Return on invested capital
9.72%
Short interest
2.59% of shares outstanding
5️⃣ Current Data Snapshot
Pricing and valuation
Metric
Value
Price
$138.02 (August 25, 2026)
Market cap
$46.32B
Shares outstanding
335.64M (+0.45% year over year)
52-week price change
-27.84%
Price-to-earnings (past year)
24.4
Price-to-earnings (next year's estimate)
14.7
PEG ratio
0.39
Enterprise value to operating cash profit
10.3
Debt to equity
3.73
Beta (volatility vs the market)
1.43
Dividend
$0.92/yr per share (0.67% yield)
Enterprise value means the market value of the shares plus debt, minus cash: what it would cost to buy the whole business outright. Operating cash profit, usually written as EBITDA, means profit before interest, taxes and accounting charges for wear and tear. It is used for power companies because those accounting charges are enormous and are not actual cash payments.
Latest quarter and guidance
Metric
June quarter 2026
Revenue
$4.02B, down 5.5%
Reported earnings per share
$0.77 (from $0.82)
Net income
$258.0M, down 7.9%
Operating cash profit
$1.767B, up over 30%
First half operating cash profit
$3.261B (from $2.589B)
First half cash from operations
$2.222B
Company forecast
Range
2026 operating cash profit
$6.8B to $7.6B, reaffirmed
2026 cash available after maintenance
$3.925B to $4.725B
2027 operating cash profit opportunity
$7.4B to $7.8B, tracking lower end
Total cash generated through 2027
More than $10B
6️⃣ Quantitative Analysis and Valuation
Analyst Ratings
Measure
Value
vs $138.02 (August 25)
Average
$225.09
+63.1%
Median
~$225
+63.1%
Consensus
Buy, 83% of 23 analysts
Reading: A 63% gap between the analyst average and the market price is extreme and should prompt caution rather than comfort. Targets have been cut by 0.46% over three months, and Wells Fargo reduced its target from $220 to $212 while keeping a positive rating. The direction is downward even though the average stays high. The median matches the average, so no single outlier is distorting it, though the full distribution across all 23 analysts has not been verified.
Valuation method, and why it differs from the usual approach
The standard method for a power company in this framework is a forward earnings multiple, with a dividend model as a floor.
This brief deviates, and the reason matters. Analyst earnings estimates for Vistra have moved substantially: the 2026 figure was revised from $9.01 to $9.40 within months, and the revenue forecast from $18.8 billion to $23.3 billion. Building purchase prices on a moving estimate produced levels that could not be defended when challenged. The company's own operating cash profit forecast is reaffirmed by management and far more stable, so that leads here, with the earnings multiple used as a cross-check.
Primary lens: enterprise value to operating cash profit
Inputs: price $138.02 (August 25, 2026), 335.64 million shares, net debt estimated at $23.5 billion (derived from the reported ratio, not read from a filing, and requiring confirmation).
Operating cash profit
9x
10x
11x
12x
$6.8B (2026 low end)
$112
$133
$153
$173
$7.2B (2026 midpoint)
$123
$145
$166
$188
$7.4B (2027 low end)
$129
$151
$173
$195
$8.1B (2027 low end plus $700M)
$147
$171
$196
$220
The extra $700 million reflects the Cogentrix purchase and the Meta contracts, which management excluded from its stated 2027 range.
Cross-check: earnings multiple on the 2026 estimate of $9.40
Multiple
Implied price
11x
$103
13x
$122
15x
$141
17x
$160
19x
$179
The current multiple is 14.7, so both methods agree the stock is close to fair value on 2026 numbers and modestly below fair value on 2027.
Dividend floor
At $0.92/yr per share with only 14% of profits paid out, a dividend-based valuation produces a figure far below the share price and provides no meaningful support. This is expected for a company that returns cash through buybacks instead, and is precisely why multiple-based methods lead here.
Stock-based compensation note
Employee share awards are not material relative to cash generation at this scale and do not require a separate adjusted valuation. To be confirmed against the annual report.
Interpretation and stress test
At $138.02 the stock sits just below the 2026 midpoint at 10x ($145) and well below the 2027 low end at 10x ($151). The base case implies modest upside. The best case requires both the Cogentrix purchase completing and investors paying a higher multiple.
Stress test: If 2027 cash profits come in at the 2026 low end of $6.8 billion, meaning no growth at all, and investors mark the multiple down to 9x, fair value is $112, an 18.9% fall from here. That is the scenario the second purchase tranche is sized for.
Buy lines: Buy under $145. Deploy double the cash under $112, where the AI premium has been removed entirely and the company is valued as an ordinary power producer.
7️⃣ Six-Category Evaluation
Category 1, Quantitative: 4.0
An earnings multiple of 14.7 on next year's estimate and a PEG ratio of 0.39 are undemanding. The enterprise multiple of 10.3 is reasonable for a contracted power producer. Cash available after maintenance of roughly $4.3 billion gives a 9.3% yield on the market value. Return on invested capital of 9.72% exceeds the cost of that capital. Why not higher: Reported revenue, net income and earnings per share all fell year over year, and management is guiding 2027 toward the lower end. Why not lower: The forecast was reaffirmed and the cash yield is genuinely high.
Category 2, Qualitative: 4.5
The revenue driver splits cleanly. Retail, roughly half of profits, is entirely independent of the AI thesis. The 3,800 megawatts under contract is locked for twenty years and survives an AI slowdown. The re-rating driver is AI-dependent: both market price capture and multiple expansion require data centre demand to persist. Operational execution has been strong, with 92 maintenance outages completed before summer and plant availability above 97% during heat waves. Why not higher: Reported earnings quality is deteriorating. Why not lower: Two named technology giants on twenty-year terms is the strongest contracted position of any name reviewed in this space.
Category 3, Classification: 4.0
Cleanly a market-exposed power producer with a nuclear fleet and integrated retail arm. Fits the Power Generation layer without ambiguity. Why not higher: Overlaps substantially with other power generation exposure. Why not lower: No classification ambiguity at all.
Category 4, Bottleneck/Moat: 4.0
The existing nuclear fleet genuinely cannot be replicated. Building new nuclear takes a decade or more and costs several times what these plants carry on the books. Operating reactors that can sign twenty-year contracts today are scarce. Why not higher: The gas and coal plants have no such advantage, and market-exposed generation is fundamentally a commodity business. Why not lower: No competitor can create a substitute for an operating reactor within the holding period.
Category 5, Risk: 3.5
The dominant risk is political action on household electricity bills, with New York's ban as a live precedent. Texas price weakness is confirmed rather than hypothetical. The 2028 gap in price protection sits inside the holding period. Partial insulation from the biggest shared risk: roughly half of profits come from retail and contracted sources, so a slowdown in AI spending caps the upside but does not break the base case. Why not higher: Political risk is real, arriving, and entirely outside management's control. Why not lower: Contracted and retail revenue provides genuine downside protection, and near-term price hedging neutralises commodity risk.
Category 6, Balance Sheet: 3.5
Investment grade at two agencies following the Fitch upgrade. Roughly 30% of shares retired since 2021, with $1.5 billion of buyback authority remaining and more than $10 billion of cash generation expected through 2027. Why not higher: Debt to equity of 3.73, net debt near $70 per share, and the Cogentrix purchase adds roughly $4 billion including assumed debt. Why not lower: Two investment grade ratings and substantial cash generation servicing the debt comfortably.
Weighted Score Calculation
Category
Score
Sector weight
Contribution
1. Quantitative
4.0
20%
0.80
2. Qualitative
4.5
15%
0.675
3. Classification
4.0
10%
0.40
4. Bottleneck/Moat
4.0
25%
1.00
5. Risk
3.5
15%
0.525
6. Balance Sheet
3.5
15%
0.525
Total
100%
3.93
Hard-fail check: No category scored 1.0 in Quantitative, Risk or Balance Sheet. No hard fail.
Interpretation tier: Medium conviction (3.0 to 3.9). The score sits at the top of the Medium band, 0.07 short of High. Risk and Balance Sheet hold it back, and both reflect genuine, measurable concerns rather than uncertainty.
8️⃣ Strategy
Primary instrument: Common shares. The cash generation and active buyback make options a poor structure, and a beta of 1.43 means the shares already provide plenty of movement.
Current position status: No position. This would be a new holding, and the larger of two positions in the power generation layer.
Scaling rule: Buy while the price is under $145, roughly the value implied by average 2026 cash profits. Deploy double the cash while under $112, where the shares would be priced as if the AI opportunity did not exist. At $138.02 (August 25) the first line is live.
Tranche
Trigger
Share of capital allocated to this name
First purchase
≤ $145 (live now)
One third
Double-down
≤ $112
Two thirds
Pause condition: Stop buying if Texas around-the-clock power prices remain below $40 per megawatt-hour through the autumn results, or if a data centre ban of the New York type is introduced in Texas or the eastern grid region. A lower price only matters if the reason for owning the shares is still intact.
Reserve expiry: If the $112 level has not been reached by the end of the first quarter of 2027 and the thesis is intact, deploy the remaining cash at whatever the market price is rather than holding it indefinitely.
Holding period: 3 to 5 years.
Reduce / exit triggers
Trigger
Action
Meta or Amazon contracts terminated or materially renegotiated
Exit
Data centre ban enacted in Texas or the eastern grid region
Reduce by half, reassess
Credit rating cut below investment grade at either agency
Reduce by half
2028 price protection falls below 60% while Texas prices stay under $40 per megawatt-hour
Reduce by half
Cogentrix completes on materially worse terms, or the share portion is repriced below $185
Reassess, no automatic action
9️⃣ Open Questions
Confirmed net debt from the June quarter filing. The $23.5 billion figure is derived from a reported ratio, not read from a filing, and it affects every price in the valuation table. (August 25, 2026)
Verify the reported forecast of 58% net income growth next year. Is this genuine, or an artifact of comparing against a depressed 2026 base? (August 25, 2026)
Exact autumn earnings date. The June quarter was reported August 7, so early November is likely but unconfirmed. Verify against company investor relations, not data aggregators. (August 25, 2026)
Latest annual employee share award expense from the annual report, to confirm the brief note in Section 6. (August 25, 2026)
The Cogentrix share consideration of 5,000,000 shares valued at $185 each, far above the current price. Confirm whether the terms adjust and what dilution results at completion. (August 25, 2026)
Full analyst distribution across all 23 analysts, to confirm the median and establish the range. (August 25, 2026)
This brief is for personal research purposes only and does not constitute investment advice. All figures should be verified against primary filings before acting.