Energy / Power · Deep dive

Vistra

The reborn corpse of history's largest LBO bust is now a $50B, ~50 GW Texas power machine signing 20-year nuclear deals with Meta and AWS — repriced by Wall Street as AI infrastructure while its batteries burned in California and PJM politicians reached for price caps.

well positioned

Vistra owns exactly what an AI-driven power shortage makes scarce — 6.4 GW of licensed nuclear and ~26 GW of dispatchable gas in the two tightest US markets — and it is converting that scarcity into 20-year hyperscaler contracts faster than regulators can cap it or the commodity cycle can turn.

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Vistra is what remains when the worst private equity deal in history finishes composting: a Texas-based competitive power producer and retailer, spun out of the Energy Future Holdings bankruptcy in October 2016, that now runs ~43.7 GW of generation (November 2025), serves about 5 million retail customers through TXU Energy and sister brands, and carries a ~$50B market cap and ~$70B enterprise value (July 29, 2026). It owns the second-largest competitive nuclear fleet in the US — 6,448 MW — and has spent the past year converting it into 20-year contracts with the two deepest pockets in the economy: AWS (1,200 MW at Comanche Peak, September 2025) and Meta (2,609 MW in PJM, January 2026). With the pending $4.7B Cogentrix acquisition adding 5.5 GW of gas, the fleet approaches 50 GW. The bull case is scarcity; the bear case is that scarcity is political, cyclical, and occasionally on fire.

Founding story

Vistra has no founder; it has a cautionary tale. In October 2007, KKR, TPG, and Goldman Sachs Capital Partners took TXU Corp private for $45B — the largest LBO ever — renaming it Energy Future Holdings. The bet: Texas power prices track natural gas, and gas would stay expensive. The shale revolution said otherwise — the ~$8B equity check was vaporized, and in April 2014 EFH filed for Chapter 11 under roughly $40B of debt. In late 2016 the competitive side of the estate — Luminant’s plants plus TXU Energy’s retail book — was spun off tax-free to first-lien creditors, who converted ~$24.4B of secured debt into 427.5 million shares and discharged ~$33.8B. The creditors named it Vistra Energy, hired merchant-power lifer Curt Morgan to run it, and listed it on the NYSE in May 2017.

The continuity is the people. Jim Burke, CEO since August 2022, joined TXU in late 2004 after Deloitte, six years in Coca-Cola’s juice business, Reliant Energy’s residential marketing, and the presidency of Gexa Energy. He ran TXU Energy — the retail cash engine — from 2005 through the LBO and the bankruptcy, then became Vistra’s CFO and COO. Burke is a retail margin operator who inherited a generation fleet and turned it into a contracts business; the corporate DNA — hedge everything, distrust leverage — is the bankruptcy talking.

How it works

Vistra is an integrated merchant: it makes electricity, sells electricity, and uses each side to hedge the other. The generation fleet (November 2025) is ~26,000 MW of natural gas, 6,448 MW of nuclear (Comanche Peak in Texas plus Beaver Valley, Davis-Besse, and Perry from the Energy Harbor deal), 4,570 MW of coal on a retirement glidepath, and 1,421 MW of solar and batteries. Output is sold three ways: forward wholesale sales and hedges, capacity payments (PJM pays generators to exist — Vistra cleared 10,566 MW at $333.44/MW-day for 2027/28), and, increasingly, long-dated corporate PPAs. The retail arm — TXU Energy, Ambit, Dynegy brands, ~5 million customers — buys that power and sells it to households and businesses at fixed rates, so when wholesale prices fall, retail margins expand, and vice versa. Add a hedging program with most expected generation sold forward through 2027, plus the IRA nuclear production tax credit as a revenue floor on the reactors, and merchant volatility starts to look like a utility annuity. Winter Storm Uri (February 2021) showed the failure mode: when plants trip and gas spikes simultaneously, the integrated hedge inverts — Vistra ate a $1.6B cash-flow hit buying gas and replacement power at panic prices.

Product and business overview

Three businesses under one roof. Generation (Luminant and the acquired fleets) is the asset base across ERCOT, PJM, ISO-NE, MISO, and California, being deliberately reshaped toward gas and nuclear: the ~$1.9B Lotus deal (seven plants, 2,600 MW, closed October 2025) and the pending $4.7B Cogentrix deal (ten plants, 5,496 MW across PJM, New England, and ERCOT) push the fleet toward ~50 GW. Retail is TXU Energy — the Texas incumbent brand — plus Ambit and Energy Harbor’s ~1M acquired accounts. The data-center franchise is the new product: the AWS PPA (September 2025) takes 1,200 MW of Comanche Peak for 20 years with 20 more in options, deliveries from Q4 2027; the Meta PPAs (January 9, 2026) take 2,609 MW in PJM — 2,176 MW from Perry and Davis-Besse starting December 2026, plus 433 MW of uprates across three plants, the largest corporate-backed nuclear uprates in US history, phasing in through 2034. Vistra Zero’s storage flagship, Moss Landing in California, is the scar tissue — see below.

Business model and pricing

Revenue books as wholesale energy and capacity in generation and as retail electricity sales, netted through aggressive forward hedging. The real price points, dated: PJM capacity at the FERC-approved cap of $333.44/MW-day for 2027/28 (July 2025 auction) — PJM itself estimated ~$530/MW-day absent the cap negotiated with Pennsylvania’s governor. Nuclear PPA terms with Meta and AWS are undisclosed but were struck to support relicensing and uprates, implying premiums above merchant curves; the nuclear PTC floors reactor revenue around the mid-$40s/MWh through 2032. TXU retail fixed plans priced in the high-teens cents/kWh in 2025-26 (Texas comparison sites), above discount rivals — TXU sells reliability and brand, not price. Capital returns are part of the model: $6.3B of buybacks since November 2021 shrank the share count ~30% to ~337M by Q1 2026, with $1.5B authorization remaining. FY2025: $17.74B revenue, $5.91B ongoing-ops adjusted EBITDA, $944M net income — the gap between the last two being depreciation, interest, and hedge mark-to-market noise.

Traction over time

The share price did more work than the EBITDA: Vistra was the single best performer in the S&P 500 in 2024 (+258%) as the market repriced merchant fleets as AI infrastructure. In 2026 the stock has sagged from records — PJM price-cap proposals and demand skepticism — leaving the market cap at ~$50.1B (July 29, 2026) against a Street mean target near $230/share, with the bear at $99 and the bull at $320 (Barchart consensus, mid-2026).

Market analysis

US electricity demand is growing for the first time in a generation: EIA (2026) projects the first four consecutive growth years since 2007, and the fight is over how much of it is real. Data-center load forecasts for 2030 range from ~65 GW of additions (Grid Strategies, 2025, the skeptic case) to 110-134 GW (Wood Mackenzie, 451 Research, 2026). Whatever the number, supply cannot respond quickly: gas turbine order books run years long, PJM interconnection queues are jammed, and the 2027/28 PJM auction cleared 6,623 MW short of the reliability requirement (July 2025). That is a sellers’ market for anyone who already owns licensed, connected, dispatchable megawatts — Vistra’s inventory. The countervailing forces: political intervention (the PJM cap already transferred ~$200/MW-day of scarcity value from generators to consumers) and the S&P Global-flagged risk (January 2026) that hyperscalers eventually bypass the grid with on-site generation and SMRs.

Competitive intel

The named set is in the sidebar. The analytical read: Constellation (~$98.7B market cap, ~60 GW post-Calpine, July 2026) owns the premium end of the story — the biggest nuclear fleet gets the first hyperscaler call, and Vistra’s Meta deal was partly proof that buyers now want a second supplier. Talen proved the model with AWS at Susquehanna and fights for the same PJM load with a leaner 13 GW. NRG is the daily knife-fight: Reliant vs. TXU in Texas retail, plus a ~$12B LS Power deal (May 2025) that rebuilt its generation flank. NextEra sells the alternative thesis — renewables-plus-storage at scale plus the Duane Arnold restart — and wins wherever speed matters less than cost and carbon accounting. Vistra’s edge over all of them is the combination: no one else pairs the second-largest competitive nuclear fleet with ~26 GW of gas and the largest Texas retail book, which is why it can sign Comanche Peak to AWS while hedging the position with its own load.

History and evolution

1912-2002: TXU and predecessors electrify north Texas; deregulation splits the business. Oct 2007: $45B KKR/TPG/Goldman LBO creates Energy Future Holdings. Apr 2014: Chapter 11. Oct 2016: TCEH spins off as creditor-owned Vistra Energy; Curt Morgan CEO. May 2017: NYSE listing. Apr 2018: Dynegy merger adds ~27 GW outside Texas. 2020-21: Moss Landing batteries (300 MW then 400 MW phases) come online. Feb 2021: Uri costs $1.6B; a Texas appeals court later ruled (March 2023) the PUC’s $9,000/MWh order exceeded its authority. Nov 2021: buyback era begins. Mar 2023: Energy Harbor deal announced ($3.43B); closes March 1, 2024 after FERC delays. 2024: best stock in the S&P 500; ~$3.2B buyout of the Vistra Vision minority closes December 2024. Jan 16, 2025: the Moss Landing 300 MW building burns — the fourth fire at the site since 2021 — forcing evacuations. Sep 2025: AWS signs for 1,200 MW of Comanche Peak. Oct 2025: Lotus gas deal closes. Jan 5, 2026: $4.7B Cogentrix deal announced. Jan 9, 2026: Meta signs for 2,609 MW. Feb 26, 2026: record FY2025 results.

What people say

The case for. The Street is constructive: consensus Buy, mean target ~$230 against a share price implying ~$50B market cap (July 2026); TIKR argued after Q1 2026 that the power-demand case “has not been priced in.” TXU Energy carries a 4.7-star rating and an NPS of 70 on comparison site 2TurnItOn (2026) — remarkable for a utility brand. Vistra scores 3.8/5 on Glassdoor with 81% willing to recommend (2026). And hyperscaler diligence is its own endorsement: Meta and Amazon each underwrote 20-year commitments to Vistra’s reactors after plant-by-plant examination.

The complaints. TXU logged 368 complaints to the Texas PUC in the trailing twelve months against an industry average of ~120 (2TurnItOn, 2026) — recurring themes are billing disputes, high plan pricing, and customer-service wait times; scale explains some of it, not all. Moss Landing is the reputational wound: residents’ suits filed from February 2025 allege Vistra used older LG NMC batteries in a flawed retrofit of a former turbine hall; litigation — including a citizens’ group running its own environmental sampling — was still working through California courts in mid-2026. The analyst bear case is threefold: PJM price-cap politics (the stock fell on 2026 proposals to cap capacity prices; TD Cowen trimmed its target to $222 in July 2026, and the low Street target sits at $99), commodity cyclicality (post-2027 hedges roll off into whatever gas curve exists then), and demand risk — if Grid Strategies’ ~65 GW view of data-center growth is right and the 125 GW views are wrong, merchant power is over-earning on a forecast.

Outlook: well positioned or at risk?

Well-positioned — because Vistra’s core asset is not a forecast, it is inventory: licensed nuclear plants and connected gas capacity in the two tightest power markets in America, at a moment when new supply is physically constrained by turbine lead times and interconnection queues into the 2030s. It is executing the right strategy for that moment: converting merchant exposure into 20-year hyperscaler contracts (3,800+ MW signed to AWS and Meta between September 2025 and January 2026), buying dispatchable gas below replacement cost (Cogentrix at roughly $850/kW versus multi-thousand-dollar new-build), and shrinking the share count 30% since 2021. The bear case deserves full weight and does not flip the verdict. Moss Landing shows the storage business’s operational tail risk and will cost years of litigation; Uri showed the hedge can invert catastrophically once a decade; the PJM cap proves that when generators win too visibly, politicians change the rules — revenue that would have cleared ~$530/MW-day cleared at $333.44 by fiat (July 2025). And if AI load disappoints, ~$70B of EV against ~$6-7B of EBITDA leaves cyclical downside the hedges only defer past 2027. But the asymmetry favors the incumbent: the PPAs already signed are contractual, not cyclical; the retail book cushions the downside scenario; and every quarter that new supply stays stuck in queues, the existing fleet compounds in value. The company that died betting on high gas prices now gets paid for owning what gas prices cannot quickly summon: firm megawatts, already built.

How a challenger would attack it

You can’t build a nuclear fleet, so you build around it. Vistra’s moat is inventory — licensed megawatts new entrants cannot summon while turbine order books run years long — so the attack is the one S&P Global already flagged: behind-the-meter generation that makes the inventory irrelevant. A hyperscaler-aligned developer selling on-site gas turbines, SMR options for the 2030s, and grid-bypass interconnection converts Vistra’s scarcity premium into a bridge product with an expiration date; every year of $333/MW-day capacity prices funds the customer’s incentive to self-supply. The second vector is retail, where the incumbent is visibly over-earning: TXU logged 368 PUC complaints against an industry average of ~120, prices fixed plans in the high-teens cents/kWh above discount rivals, and sells brand rather than price. A retail attacker pairing transparent pricing with smart-device load-shifting attacks the cash engine that funds Vistra’s hedge book — and in ERCOT, switching is a web form. Third, the political flank: the PJM cap transferred ~$200/MW-day from generators to consumers by negotiation with one governor. A challenger structured as a consumer-aligned aggregator — demand response, distributed batteries, monetizing load flexibility — profits from every intervention that caps Vistra’s scarcity rents, effectively shorting the incumbent’s regulatory exposure while Moss Landing litigation keeps the storage flank reputationally pinned.

Same playbook, new buyer

Vistra’s real playbook — buy distressed or unloved dispatchable assets below replacement cost, hedge them with a captive retail book, then convert scarcity into long-dated contracts — is being executed by every large IPP for the same buyer: US hyperscalers. The uncontested version changes the buyer, not the asset class. Industrial electrification is the nearer shift — steelmakers, hydrogen producers, and desalination need firm power on 10-20 year horizons just like data centers, but no one is packaging existing fleets into industrial PPAs because AI demand is louder; if Grid Strategies’ skeptic case (~65 GW, not 125) proves right, whoever built the industrial-offtake channel owns the fallback demand while rivals over-earn on a forecast. Vistra won’t pursue it seriously while Meta-grade counterparties pay premiums above merchant curves. The second shift is geographic: the integrated merchant-plus-retail model is an ERCOT artifact, and deregulating markets abroad — where AI load growth meets thin capacity markets — lack any operator running the generation-hedged-by-retail machine Burke’s team perfected through a bankruptcy. Vistra’s DNA — distrust of leverage, Texas-centric retail brands, US regulatory fluency — keeps it home, leaving the export version of its own model unclaimed.

Sources and further reading

• Vistra Reports Fourth Quarter and Full-Year 2025 Results — Vistra IR, February 26, 2026
• Vistra and Meta Announce Agreements to Support Nuclear Plants in PJM — Vistra IR, January 9, 2026
• Vistra to buy Cogentrix Energy’s 5.5 GW of gas plants — Utility Dive, January 2026
• Vistra secures long-term nuclear PPA from Comanche Peak nuclear plant — Power Engineering, September 2025
• PJM capacity prices hit price cap, reserve shortfall grows — Utility Dive, July 2025
• Wave of litigation follows January fire at Vistra Moss Landing battery facility — Local News Matters, May 22, 2025
• Vistra: Texas freeze caused $1.6B negative cash flow impact — Power Engineering, April 2021
• Energy Future Holdings: the second-largest public utility filing ever — Louisiana Law Review, June 2018
• With FERC approval in hand, Vistra poised to close on $6.3B Energy Harbor buy — Utility Dive, February 2024
• Vistra (VST) Statistics & Valuation — stockanalysis.com, July 29, 2026
• Vistra Reports First Quarter 2026 Results — Vistra IR, May 7, 2026