Vistra Corp

Vistra is one of the largest US power generators and retailers — nuclear, natural gas, solar, and storage. Owns Comanche Peak (two 1,200 MW nuclear units in Texas) and extensive natural gas peaker fleet. Sells power into competitive wholesale markets (ERCOT, PJM) and has signed data-center/AI power supply agreements.

Vistra is a large integrated US power generator and retail supplier with nuclear, gas, solar, and storage assets. The thesis here is not generic utilities exposure; it is scarce dispatchable and carbon-free generation capacity that hyperscalers can contract for AI/data-center load.

Vistra is a large merchant power generator and retailer with nuclear, gas, solar, and storage assets. The thesis is not generic utility exposure; it is scarce licensed baseload and dispatchable capacity that hyperscalers can contract for AI/data-center load.

Vistra is a large US merchant power generator and retailer with nuclear, gas, solar, and storage assets. The Physical AI thesis is not generic utilities exposure; it is scarce, licensed, dispatchable nuclear MW that hyperscalers can contract for AI/data-center load.

Vistra is a large US merchant power generator and retailer with nuclear, gas, solar, and storage assets. The investable bottleneck is not broad utility exposure; it is scarce licensed nuclear and dispatchable generation capacity that hyperscalers can contract for AI/data-center load.

Q2 2026 operating revenues were $4.017B versus $4.250B in Q2 2025; operating income was $553M versus $515M; net income attributable to Vistra was $305M versus $327M; Ongoing Operations Adjusted EBITDA was $1.767B versus $1.349B. The company reaffirmed 2026 Ongoing Operations Adjusted EBITDA guidance of $6.8B-$7.6B and Ongoing Operations Adjusted FCFbG guidance of $3.925B-$4.725B.

Vistra is a US merchant power generator and retailer whose investable bottleneck is scarce, licensed nuclear and dispatchable generation capacity sold into long-duration hyperscaler power contracts.

Vistra is a US merchant power generator/retailer whose investable bottleneck is scarce, licensed nuclear and dispatchable generation capacity sold into long-duration hyperscaler power contracts. The Q2 10-Q confirms the fleet is being progressively contracted (AWS, Meta) while Vistra deploys balance-sheet capital into the KKR Helix fund to capture AI data-center buildout economics.

• monetization model: Integrated retail electricity + wholesale power generation: retail electricity sales, wholesale energy/capacity/ancillary services, RECs, hedges; not subscription/SaaS. (Vistra FY2024 10-K Item 1; EDGAR CIK 0001692819, sec.gov ; accession UNKNOWN here.)
• revenue mix: Segments: Retail, Texas, East, West, Asset Closure, Corporate/Other (FY2024 10-K Note 15). Retail largest; generation segments Texas/East/West. Exact FY2024 % UNKNOWN here. FY2023 10-K Note 15: Retail ~67%, Texas ~32%, East ~23%, West ~10% of consolidated revenue; intersegment eliminations make shares sum >100% (sec.gov , accession 0001692819-24-000008).
• who pays: Residential, commercial, industrial retail customers; wholesale counterparties/utilities/co-ops/municipalities. (10-K Item 1).
• margin_value_capture: Integrated generation-to-retail margin, capacity/ancillary, hedging; gross/operating margin not disclosed as a single metric. Value accrues in Retail and generation segments. (10-K Item 7, Note 15).

• customer classes: Residential retail customers, commercial/industrial customers, wholesale counterparties; Utility - Grid Operators, Industrial Infrastructure Owners; data-center/hyperscaler demand sits inside commercial/industrial but not separately quantified in 10-K. (10-K Item 1).

Vistra is a vertically integrated power generation and competitive retail electricity business. Its model is best understood as:

1. Own and operate dispatchable generation — natural gas, coal, nuclear, solar, and battery storage.
2. Sell generation output into wholesale power markets — principally ERCOT, PJM, ISO-NE, MISO, and CAISO.
3. Sell electricity and natural gas directly to residential and commercial customers — mainly through Texas retail brands.
4. Use retail and generation together to hedge merchant exposure — the retail book gives a natural hedge for part of the generation fleet.

This structure is documented in [VST-10K-2023], Item 1, “Business,” and Item 7, “Management’s Discussion and Analysis of Financial Condition and Results of Operations.”

Vistra’s reportable segments per the FY2023 10-K are:

Source: [VST-10K-2023], Item 8, “Reportable Segments.”

I will not publish a falsely precise segment percentage table. The exact segment profit/revenue reconciliation requires the segment footnote in [VST-10K-2023], because segment revenues include inter-segment sales and consolidated revenue eliminates those inter-segment flows. That means a simple “Retail = X% of total revenue” from a sell-side summary can be misleading.

That said, the broad structure is clear from the filing:

• Retail is the largest revenue segment by absolute external sales.
• Texas generation is the second-largest revenue segment.
• East generation is the third-largest.
• West, Sunset, and Asset Closure are smaller, though Sunset and Asset Closure carry material liabilities and decommissioning obligations.

Vistra monetizes electricity in several distinct ways:

• Wholesale energy sales: sells electricity in RTO/ISO real-time and day-ahead markets.
• Capacity payments: in organized markets such as PJM and ISO-NE, earns capacity revenue for being available.
• Ancillary services: provides frequency regulation, spinning reserves, and voltage support.
• Retail electric plans: sells fixed-rate, variable-rate, and indexed plans to residential and commercial customers.
• Natural gas retail: sells retail natural gas, mainly through acquired retail brands.
• Renewable/nuclear tax credits: after the Inflation Reduction Act, monetizes nuclear production tax credits and renewable credits.
• Battery storage arbitrage and ancillary services: charges during low-price periods, discharges during high-price periods.

Source: [VST-10K-2023], Item 1 and Item 7.

The payers are broad and include:

• Texas residential and commercial customers through retail brands.
• Other competitive retail customers in the Northeast/Midwest.
• RTO/ISO markets: ERCOT, PJM, ISO-NE, MISO, and CAISO.
• Load-serving entities, municipalities, cooperatives, and other utilities that purchase capacity and energy.
• Counterparties in tolling, heat-rate, and full-requirements contracts.

Source: [VST-10K-2023], Item 1.

The FY2023 10-K does not disclose a named top end-customer that would account for a material revenue concentration. The named counterparties are primarily RTO/ISO market operators and financial settlement counterparties, not individual end-use customers. This is important: any source claiming a specific major customer relationship should point to a contract filing or regulatory disclosure. I did not find a named top customer in [VST-10K-2023].

From [VST-10K-2023] Item 1:

• Residential customers — primarily in Texas under TXU Energy and Ambit Energy.
• Small and large commercial/industrial customers — served through retail and wholesale channels.
• Governmental, municipal, and cooperative load-serving entities — as wholesale purchasers.
• Other utilities and load-serving entities — as counterparties for capacity and energy sales.
• RTO/ISO markets — as the central settlement and credit counterparty.

The buyer archetypes are:

1. Texas residential household — buys fixed-rate, variable-rate, or indexed electric plans.
2. Small commercial business — buys retail electricity and natural gas.
3. Large industrial customer — buys energy via structured retail or wholesale contracts.
4. RTO/ISO load-serving entity — buys capacity, energy, and ancillary services.
5. Tax-equity investor or credit buyer — participates in tax-credit transfer transactions.

None disclosed in the FY2023 10-K as material top customers. The named counterparties are market operators and financing counterparties. I do not invent end-customer names.

FY25 Ongoing Operations Adj. EBITDA $5.912B (+$269M vs FY24); Net Income reduced $(1,868)M driven by unrealized mark-to-market losses on hedges; power generation is the core; announced Cogentrix acquisition (~5,500 MW gas capacity); also closed Lotus Infrastructure ~2,600 MW acquisition.

Q1 2026 revenue from contracts with customers was $5.001B (+17.7% YoY vs $4.250B in Q1 2025); operating income was $1.499B versus a $120M loss a year ago; net income was $1.029B. TTM revenue is $19.45B, gross margin 42.7%, EBITDA $6.77B, market cap $52.41B, and forward P/E 14.83.

Q1 2026 revenue from contracts with customers was $5.001B (+17.7% YoY vs $4.250B in Q1 2025); operating income was $1.499B, implying a 30.0% operating margin; net income was $1.029B, implying a 20.6% net margin. Cash and cash equivalents were $634M and long-term debt was $19.163B (SEC companyfacts / Q1 2026 10-Q).

Q1 2026 revenue from contracts with customers was $5.001B (+17.7% YoY vs $4.250B in Q1 2025); operating income was $1.499B (30.0% margin vs -2.8% a year ago); net income was $1.029B (20.6% margin vs -6.3% a year ago). Cash and cash equivalents were $634M and long-term debt was $19.163B.

Q2 2026 operating revenues were $4.017B, down 5.5% YoY from $4.250B; operating income was $553M, implying a 13.8% margin; net income attributable to Vistra was $305M, implying a 7.6% margin; Ongoing Operations Adjusted EBITDA was $1.767B, up 30.9% YoY and equal to a 44.0% EBITDA margin. Latest filed balance sheet data (Q1 2026 10-Q) showed cash and cash equivalents of $634M and long-term debt of $19.163B.

Q1 2026 revenue from contracts with customers was $5.001B; operating income was $1.499B; net income was $1.029B; cash and cash equivalents were $634M; long-term debt was $19.163B. On the supply side, Vistra owns six nuclear units totaling 6,448 MW. The named AWS tranche is 1,200 MW and the Meta tranche is 2,609 MW, so contracted hyperscaler nuclear capacity is 3,809 MW — about 59.1% of the nuclear fleet.

Q2 2026 operating revenues were $4.017B (-5.5% YoY vs $4.250B), but the underlying contracted revenue grew sharply — see the mark-to-market reconciliation below. H1 2026 operating income was $2.052B vs $395M a year ago (+419% YoY); H1 net income before tax was $1.639B vs a -$41M loss a year ago. Q2-end balance sheet: cash $435M (down from $634M at Q1-end), long-term debt $19.595B (up from $19.163B), total assets $42.603B, stockholders' equity $5.482B, net PP&E $19.944B.

AI hyperscalers signing multi-year power purchase agreements with nuclear operators for zero-carbon baseload — Vistra's Comanche Peak = direct beneficiary. AWS deal + Meta deal confirmed.

The company is benefiting from AI/data-center load growth because large-scale electricity consumers need long-duration baseload and peaking power. Vistra’s nuclear fleet and broad dispatchable portfolio are exactly the kind of supply-side bottleneck the physical-AI buildout must secure.

AI and hyperscaler load growth need long-duration power. Vistra's nuclear fleet and ERCOT/PJM interconnection are the scarce assets; the company is monetizing that scarcity by selling block power to very large electricity consumers.

The bottleneck is long-duration, carbon-free, grid-connected nuclear capacity. Vistra monetizes that scarcity by selling multi-decade power blocks to hyperscalers that cannot self-supply equivalent baseload on their own timeline.

AI hyperscalers need long-duration, carbon-free baseload; Vistra monetizes that scarcity through multi-year power blocks and preferred-power relationships. The thesis is driven by contractable MW, not brand.

Each hyperscaler PPA converts scarce baseload MW into long-dated contracted cash flow. The moat is the mismatch between 20-year contracting needs and 15+ year replacement timelines for nuclear buildouts.

The moat is the mismatch between hyperscalers' need for 20-year carbon-free baseload and a 15+ year nuclear replacement timeline. The Q2 10-Q sharpens this into two hard numbers: (1) a record PJM capacity clearing price, and (2) a $881M mark-to-market swing that is hiding +17% underlying contracted-revenue growth.

• 800 MW nuclear power PPA signed with Amazon Web Services at Comanche Peak [SEC 8-K EX-99.1 2026-02-26]
• Nuclear energy + capacity + uprates signed with Meta at PJM nuclear facilities [SEC 8-K EX-99.1 2026-02-26]
• Cogentrix acquisition ~5,500 MW natural gas capacity announced; Lotus ~2,600 MW gas capacity closed — portfolio expanded significantly [8-K 2026-02-26]
• FY25 Ongoing EBITDA $5.912B — confirms generation scale and profitability
• CROSS-MENTIONS: AWS (Amazon) as named customer for Comanche Peak nuclear capacity
• Q1 2026 revenue from contracts with customers rose to $5.001B (+17.7% YoY) and operating income improved to $1.499B from a $120M loss a year earlier [SEC companyfacts / 10-Q filed 2026-05-08]
• The 10-Q says demand from data centers and other large-scale electricity consumers is contributing to projected fast-paced load growth in the regions Vistra serves, and that the company continues to engage with counterparties on long-term power sales from its generation fleet [10-Q 2026-05-08]
• July 10, 2026 8-K: Vistra amended its receivables purchase agreement; June 24, 2026 8-K: Vistra amended its credit agreement and commodity-linked credit agreement, signaling active balance-sheet / liquidity management [8-Ks 2026-07-10 and 2026-06-24]
• 2025 10-K, Item 1: Vistra owns and operates six nuclear units totaling 6,448 MW; Comanche Peak Unit 1 is 1,200 MW and licenses to 2050, Comanche Peak Unit 2 is 1,200 MW and licenses to 2053; the filing says nuclear units are generally operated at full capacity.
• 2026-02-26 8-K EX-99.1: Vistra announced an 800 MW nuclear power PPA with Amazon Web Services at Comanche Peak and nuclear energy, capacity, and uprates with Meta at PJM nuclear facilities.
• 2026-03-31 10-Q, Business Environment and Outlook: large-scale data centers are contributing to projected fast-paced load growth; Vistra says it is engaging counterparties on potential long-term power sales from its generation facilities.
• 2026-07-10 8-K: amended the receivables purchase agreement, increasing committed purchasers from $1.1B to $1.25B and extending the term to July 9, 2027; 2026-06-24 8-K amended credit agreements, signaling active liquidity management.
• 2025 10-K, Item 1: Vistra owns and operates six nuclear generation units totaling 6,448 MW; Comanche Peak Unit 1 is 1,200 MW and licenses to 2050, and Comanche Peak Unit 2 is 1,200 MW and licenses to 2053; the filing says nuclear units are generally operated at full capacity.
• 2025 10-K: Vistra announced a 20-year AWS PPA to supply 1,200 MW of carbon-free power from Comanche Peak, with delivery beginning in 4Q27 and ramping to full capacity by 2032.
• 2026 Q1 10-Q: data-center demand and other large-scale electricity consumers are contributing to projected fast-paced load growth in the regions Vistra serves; the company is actively discussing long-term sales from its generation fleet and is pursuing nuclear uprates and other capacity expansions.
• 2026 Q1 companyfacts: revenue $5.001B, operating income $1.499B, net income $1.029B, cash $634M, long-term debt $19.163B.

• 1,200 MW × 8,760 hours = 10,512,000 MWh/year = 10.512 TWh/year at full output.
• At a 90% capacity factor, that is 9.461 TWh/year; at 85% it is 8.935 TWh/year.
• The same 1,200 MW equals 50.0% of Comanche Peak and 18.6% of the full 6,448 MW nuclear fleet.
• For every $1/MWh change in realized pricing, AWS tranche annual revenue changes by $10.512M at full output or $9.461M at 90% CF; over 20 years at 90% CF, that is $189.2M of cumulative revenue per $1/MWh.

A 1,200 MW tranche at nameplate output implies 10.512 TWh/year (1,200 × 8,760 hours). At a 90% capacity factor, that is 9.461 TWh/year, so every $1/MWh change in realized pricing moves annual revenue by about $9.46M. That same tranche is 50.0% of Comanche Peak's 2,400 MW nameplate and 18.6% of Vistra's 6,448 MW nuclear fleet, which shows why each signed hyperscaler block meaningfully re-prices the asset base.

• The 2025 10-K says Comanche Peak Unit 1 and 2 are each 1,200 MW, are licensed to 2050 and 2053 respectively, and that nuclear units are generally operated at full capacity.
• The same 10-K announced a 20-year AWS PPA for up to 1,200 MW from Comanche Peak, with deliveries beginning in Q4 2027 and ramping to full capacity by 2032.
• The Q1 2026 10-Q says data-center load growth is accelerating power demand in Vistra's regions and that Vistra continues to engage counterparties on long-term sales from its generation fleet.
• Combined AWS + Meta hyperscaler PPAs total 3,809 MW, or about 59.1% of the nuclear fleet, leaving a finite residual base of scarcity.

1,200 MW × 8,760 = 10.512 TWh/year at nameplate output. Relative to the 6,448 MW nuclear fleet, AWS consumes 18.6%; relative to Comanche Peak, it is 50.0%. At 3,809 MW contracted to hyperscalers, Vistra has already pre-sold about 59.1% of its nuclear fleet. That means the remaining uncontracted nuclear capacity is the scarce optionality, not an unlimited commodity supply.

• Underlying contracted revenue (revenue from contracts with customers, excl. mark-to-market) was $4.401B in Q2 2026 vs $3.753B in Q2 2025 — +17.3% YoY — while GAAP "operating revenues" printed -5.5%. The divergence is a $881M mark-to-market swing (a +$497M gain in Q2 2025 flipping to a -$384M drag in Q2 2026), not an operating decline.
• PJM Capacity Auction for planning year 2028-2029 (results July 2026): Vistra cleared 10,924 MW at $325.00/MW-day across all seven zones — a historically strong clearing price that monetizes the data-center-driven scarcity.
• KKR Helix fund (Note 7, equity method): Vistra is a founding investor in KKR Helix C L.P., an open-ended fund managed by KKR, committing up to $1.0B = $500M initial + $500M additional subject to power-supply milestones (or at Vistra's election). Mandate spans hyperscale data-center development, baseload/flexible generation, T&D, and fiber/connectivity.
• Meta PPA plant-level detail: 20-year PPAs for 2,609 MW total = Perry 1,268 MW + Davis-Besse 908 MW (operating, 2,176 MW) plus 433 MW of uprates (Perry 213 MW, Davis-Besse 80 MW, Beaver Valley 140 MW). Operating energy/capacity begins delivering late 2026 (full by end-2027); uprates begin 2031 (full by end-2034), with uprate capex spanning 2026-2034 (<20% of spend by end-2028).
• AWS PPA: 20-year (extendable 20 more), 1,200 MW carbon-free from Comanche Peak, delivery Q4 2027, ramp to full by 2032.
• Portfolio actions: sale of Casco Bay / Beaver Falls / Syracuse gas plants (754 MW, ISO-NE + NYISO, East segment) expected to close H2 2026; FERC approved the Cogentrix acquisition (from 8-K 08-07).
• Dorrell antitrust litigation (nuclear compensation) was dismissed without prejudice in August 2026 (Maryland District Court).

• Mark-to-market reconciliation: contracted Q2 revenue $4.401B - GAAP $4.017B = -$384M MTM drag in Q2 2026; Q2 2025 was +$497M. The $881M swing fully explains why reported revenue fell 5.5% while contracted sales grew 17.3%.
• PJM capacity revenue: $325/MW-day × 10,924 MW × 365 days ≈ $1.30B/year of capacity revenue alone in the 2028-2029 planning year.
• Contracted fleet: (1,200 + 2,176) / 6,448 = 52.4% of nuclear fleet contracted on an operating basis; 3,809 / 6,448 = 59.1% including uprates.
• Uprate energy: 433 MW × 8,760 h × 90% CF ≈ 3.41 TWh/year of incremental carbon-free energy once ramped (by 2034).
• Helix commitment vs balance sheet: the $500M initial commitment exceeds Q2-end cash of $435M; long-term debt rose $432M in Q2 ($19.163B → $19.595B) — the "cost of the moat" is being funded on the balance sheet.