Vistra: AI Makes Power Plants More Valuable, But a 31% Drop Still Isn't Cheap

I am bullish on Vistra's (VST) profit growth over the next year, but remain more discerning regarding its current stock price. Based on the earnings forecasts and valuation assumptions in this article, my 12-month reference target price is $166, which is approximately 9% above the closing price of $151.72 on September 8, 2026. Although the stock price has pulled back about 31% from its 52-week high, the remaining upside remains limited even when factoring in growth from signed contracts and acquisitions.

Source: TradingKey

VST's opportunities are very tangible: its existing power plants can generate more profit, and long-term nuclear power contracts make a portion of its revenue more predictable. The company is indeed improving. My disagreement lies in the price—over the coming year, even if earnings grow as expected, it may not necessarily translate into high stock returns.

Generating and Selling Its Own Power: VST's Profits Depend on More Than Just Electricity Prices

VST operates two complementary businesses: power generation and retail. The generation division operates power plants, selling electricity into the wholesale market or to contracted customers; the retail division provides electricity plans to households and businesses, charging based on contracts. The company possesses approximately 44 GW of generation capacity and serves about 5 million retail customers. GW stands for gigawatt; 1 GW equals 1,000,000 kilowatts, used to measure a power plant's generation capacity.

Natural gas is the primary driver of VST's scale, while nuclear power's importance exceeds its share of installed capacity. Installed capacity represents how much power a plant can generate, whereas actual generation also depends on how long it operates throughout the year. Nuclear power plants run for longer periods; thus, roughly 15% of installed capacity contributes about 24% of actual power generation.

Power generation uses the company's disclosed portfolio basis, including full-year output from the acquired Lotus assets.

Profits from the generation business primarily depend on how much electricity sales revenue exceeds generation costs. Gas-fired plants are particularly sensitive to natural gas prices: if electricity sales revenue increases by 10 yuan while fuel costs rise by 7 yuan simultaneously, the additional profit margin is only 3 yuan. Nuclear power costs do not share the same sensitivity to natural gas prices, making it easier to retain gains from such electricity price increases. Coal power can also benefit from higher electricity sales prices, but still bears fuel, maintenance, and environmental compliance costs.

The retail business, meanwhile, helps buffer the company against wild swings. After households sign up for a one-year fixed-rate electricity plan, VST typically must supply electricity at the agreed price during the contract period. When wholesale electricity prices rise, the generation division sells at higher prices; however, procurement costs for the retail division that were not locked in advance will rise, squeezing profits. When electricity prices drop, the effect is reversed. The parts where conditions like volume and time periods match between both sides can offset each other; after customer contracts expire, retail prices are readjusted.

Therefore, VST's advantage lies in having both power plants and end customers. The generation business provides upside opportunities, while the retail business mitigates part of the price volatility. Evaluating the company requires looking at not only how high electricity can be sold for, but also fuel costs and how much of the sales price has already been locked in.

The Tighter the Power Supply, the More Valuable Existing Power Plants Become

I have greater confidence in the earnings support from the eastern U.S. market, while remaining relatively cautious about Texas. Together, these two regions account for over 80% of VST's power generation: Texas accounts for about 42%, and the PJM market in the Eastern and Midwestern U.S. accounts for about 40%. PJM operates the regional grid and wholesale electricity market, while the primary electricity market in Texas is operated by ERCOT.

PJM is already experiencing insufficient backup power supply, making existing power plants harder to replace. The power grid cannot prepare power sources based solely on normal electricity consumption; it must also reserve margin for extreme heat, cold waves, and sudden plant outages. PJM conducts auctions in advance to purchase commitments from power plants and other resources to 'provide power at critical moments'—known in the industry as capacity procurement. By fulfilling this commitment, power plants earn capacity revenue, and subsequently collect electricity fees upon actual generation.

Currently, this power supply guarantee has not been sufficiently procured. For the period from June 2028 to May 2029, the power supply capacity secured by PJM is approximately 6.8 GW short of the reliability standard requirement. The implication here is that the emergency margin intended beyond expected electricity consumption is insufficient, leaving the system more vulnerable during extreme weather or equipment failures. What is lacking is safety margin, rather than an active daily deficit of 6.8 GW in power supply.

Shortfalls in two consecutive delivery years indicate that new power supply resources have not kept pace with readiness requirements. Data centers submitting applications only represent potential future electricity use; capacity procurement, on the other hand, requires resource providers to make concrete commitments to guarantee power supply in designated delivery years. The latter still fell short of its targets, providing a more concrete rationale for 'tight power supply.' This is also the main reason why I believe VST's existing power plants can maintain solid returns.

The situation in Texas is somewhat different. The North American Electric Reliability Corporation (NERC), the entity that assesses North American power grid reliability, noted in its '2026 Summer Reliability Assessment' that energy storage expansion has reduced power supply risks during Texas's evening peak. Batteries can charge when power is abundant and discharge after the sun sets and solar output declines. Consequently, I do not factor sustained sharp increases in Texas electricity prices into my forecasts. While demand continues to grow, energy storage expansion is reducing the probability of extreme power shortages and will also temper the windfall profits VST's gas-fired plants earn from extreme price spikes.

Locking in Prices Early Allows a Good Business to Last Longer

Whether VST can achieve a higher valuation hinges on how long improved power sales terms can be sustained. If high electricity prices last for only a single year, VST remains a cyclical business; only when multi-year future revenues become stable and predictable does earnings quality truly transform. The company is currently using short-term advance pricing and long-term customer contracts to alter VST's historically cyclical nature.

Power sales prices over the next year already offer high visibility. As of August 3, 2026, VST's hedge ratio for its projected 2027 power generation stood at approximately 94%. Hedging can be simply understood as covering part of the risk of future power price fluctuations through advance contracting and other methods. It mitigates the impact of price drops while also capping windfall gains from subsequent price increases.

This means that 2027 growth should be evaluated based on locked-in sales terms rather than how high spot electricity prices might surge. Benefits from prior increases in electricity demand may already be reflected in these contracts. Moving forward, what matters more is whether the company can continue locking in generation profits for subsequent years under favorable terms. Nevertheless, plant outages and cost changes will still affect outcomes, as advance pricing primarily addresses price risk.

A larger transformation stems from 20-year nuclear power contracts driven by AI. The key value of nuclear contracts is enabling VST to boost returns on existing power plants while reducing future revenue dependence on market power prices. AI data centers require long-term, stable power supplies, and tech companies seek to align with low-carbon targets, making them willing to sign long-term power purchase agreements in advance. Signed contracts with Amazon and Meta involve roughly 3.38 GW of existing nuclear capacity, exceeding half of VST's existing nuclear fleet. This is already sufficient to reshape the revenue structure of a substantial portion of its nuclear business.

What I value more is that this growth relies primarily on improving returns from legacy power plants without requiring new capital expenditures to build nuclear plants of equivalent scale. The original electricity was already being sold; the increment brought by new contracts represents additional cash flow from superior sales terms, while a 20-year customer commitment reduces the uncertainty of renegotiations upon short-term contract expirations. For valuation purposes, what warrants a premium is that the exact same set of assets has the opportunity to earn more with greater revenue predictability, rather than simply having '20 years' written on the contract term.

Profit increments over the next year will primarily depend on Meta, while Amazon will support growth further down the line. Deliveries under Meta's contract for existing units begin in Q4 2026 and are expected to reach full scale by Q4 2027, meaning 2027 will only reflect a partial annual contribution. The Amazon contract starts in Q4 2027 and is not expected to be fully delivered until 2032. My assessment is that these contracts will incrementally raise VST's earnings rather than producing a one-time step-up; in the short term, execution depends on Meta, while in the long run, it hinges on whether more nuclear capacity can replicate this sales model.

In addition, regulatory dynamics make long-term contracting even more noteworthy. High prices earned by generators can ultimately translate into hefty electric bills for households and businesses. Regulators must balance controlling the rate burden while maintaining sufficient returns to incentivize new power plant construction. The U.S. Federal Energy Regulatory Commission (FERC) has approved PJM's extension of price caps and floors in capacity auctions through the 2029/2030 delivery year, curbing extreme price volatility. While VST can benefit from tight power supply, I would not assume that worsening shortages allow it to raise prices indefinitely.

2027 Earnings Expected to Grow 17%, Driven Primarily by Three Sources

I project VST will achieve approximately $8.4 billion in adjusted EBITDA in 2027, representing a roughly 17% increase over the midpoint of the company's 2026 guidance of $7.2 billion. Growth will mainly stem from core business improvements, the acquisition of Cogentrix gas-fired power plants, and the Meta contract. EBITDA stands for earnings before interest, taxes, depreciation, and amortization; the adjusted metric further excludes items such as unrealized hedging gains and losses.

The core business and the acquisition form the main pillars of this forecast. The company's baseline earnings range for 2027 is $7.4 billion to $7.8 billion, which excludes Cogentrix and Meta. Cogentrix's contribution of approximately $550 million is estimated based on the transaction price and corresponding earnings multiple disclosed by the company, assuming the transaction closes before the end of 2026 and contributes full-year earnings in 2027. Meta will still be ramping up deliveries in 2027, so only a partial contribution estimated at around $250 million is included. Contract pricing and quarterly delivery volumes have not been publicly disclosed, so actual increments will depend on delivery progress.

However, it is worth noting that nearly half of the incremental earnings comes from the acquisition. Out of the roughly $1.2 billion earnings increment, Cogentrix contributes about $550 million; acquiring these profits requires cash payments, debt assumption, and share issuance. Consequently, the 17% operational earnings growth cannot be directly equated to earnings-per-share (EPS) growth, as financing costs and share count expansion will also impact shareholder returns.

VST Deserves a Premium Over the Past, but Current Valuation Is Approaching Fair Level

VST's valuation midpoint has risen significantly compared to the past, but I believe 9x–10x is closer to a fair level today. Calculated as year-end enterprise value divided by adjusted EBITDA from continuing operations for that year, multiples were roughly 6.5x, 5.1x, and 6.4x in 2019, 2020, and 2023, respectively; they rose to about 12x in 2024 and around 12.5x in 2025. Over the past two years, the market has markedly altered how it prices VST.

I believe past valuations of 5x–7x look low by today's standards, but multiples above 12x are also unlikely to become a long-term norm. VST now benefits from two strong pillars it previously lacked: first, 20-year long-term nuclear contracts with companies like Meta and Amazon, which provide greater visibility for future revenues of certain existing power plants; second, its retail business and high hedge ratios, which protect overall profits from the severe volatility seen by pure wholesale power generators. The company expects that as signed contracts gradually materialize, the retail business and long-term contracts could contribute nearly half of adjusted EBITDA in the future. These shifts are sufficient to justify a higher normalized valuation for VST than in the past.

However, I would not grant VST an excessively high valuation based on this alone. Long-term contracts are currently concentrated mainly in nuclear power, while large asset holdings in natural gas and coal remain exposed to power prices, fuel costs, and supply-demand dynamics. Furthermore, part of the 2027 earnings growth stems from the Cogentrix acquisition, which fundamentally involves spending money to buy more assets rather than purely expanding returns from existing power plants. Therefore, I prefer to view VST as an integrated power generator with improving earnings stability, rather than a pure long-term contracted infrastructure firm that has completed its transformation.

Based on this assessment, I apply a 9.5x valuation multiple to 2027 adjusted EBITDA. This is noticeably higher than VST's historical normalized level of roughly 5x–7x, yet lower than the valuations exceeding 11x or even 14x at the height of the AI power trade—effectively recognizing that the business model has improved, without prematurely pricing in all future nuclear contracts and power price increases.

Applying a 9.5x valuation multiple to my projected 2027 adjusted EBITDA of approximately $8.4 billion, and incorporating post-acquisition changes in debt, preferred stock, and share count, I estimate VST's fair value over the next 12 months to be approximately $166.

This price target also explains why I am bullish on VST's earnings yet hesitate to issue a more aggressive recommendation near current levels. The question is no longer whether VST deserves a higher valuation than in the past, but whether it merits further re-rating above 9.5x. Currently signed long-term contracts are sufficient to support a higher valuation midpoint, but not enough to convince me that 11x–12x should become the new normal.

For the stock to rise substantially above $166, new catalysts need to emerge: signing high-quality long-term contracts for the remaining nuclear capacity, or continuously locking in generation profits at higher levels over the coming years. Otherwise, even if 2027 earnings grow to around $8.4 billion, the current stock price has already priced in a significant portion of these improvements in advance.

My conclusion is straightforward: the era of VST trading at 5x–7x is likely over, and 9x–10x represents a more reasonable valuation midpoint today; however, buying at current price levels means the market has already acknowledged this upgrade, leaving risk-reward odds that are no longer particularly attractive.

This article is for research and informational purposes only and does not constitute investment advice.