The MoatNarrow moat

Vistra (VST) — moat facet

Real assets, a genuine structural advantage, and a return on capital that has cleared its hurdle in two of nine years.

Electricity is the purest commodity in this collection. A megawatt-hour from Vistra's plant is identical to a megawatt-hour from the plant next door, the price is set by whichever generator is last to be dispatched, and no customer has ever paid a premium for the brand on the electron. On the face of it this should be Barrick with cooling towers.

Return on invested capital vs 8% cost of capitalWACC ~8%1.0%20172.3%20187.0%20194.9%20209.5%202313.8%20245.6%20252021 (-5.4%) and 2022 (-4.2%) omitted - Winter Storm Uri and hedging marks
Two years above the hurdle out of nine, with the two negative years left off the axis. This is why the rating is narrow rather than wide.

It is better than that, for three specific reasons, and the reasons are worth separating from the AI story that currently surrounds the company.

The first is that some of the assets genuinely cannot be reproduced. Vistra owns six nuclear reactors at four sites totalling 6,448 megawatts, with operating licences running to dates between 2036 and 20531. Nobody is going to build a competing one. Vistra did not build these either — it bought 4,048 megawatts of them for $3.1 billion2, which works out near $766 a kilowatt for plant that is already licensed, already staffed and already connected. That is a different kind of asset from a gas turbine anyone with capital can order.

The second is the retail book. Five million customers3 is not a moat in the switching-cost sense — a Texan can change electricity provider in an afternoon — but it is a structural advantage of a subtler kind. A generator with no customers has to sell into a spot market and hedge the price. A generator that owns a retail business has a natural offset: when wholesale prices spike, the plants earn more and the retail arm pays more, and the two partly cancel. Vistra's own description is that the integrated model lets it structure products in a way a stand-alone retailer cannot4. The $8.5 billion of intersegment sales5 is what that looks like in the accounts.

The third is location. Vistra's fleet sits in ERCOT, which manages about 83,707 megawatts of Texas peak demand for roughly 27 million customers6, and in PJM, which manages about 160,709 megawatts for about 67 million7. These are the two markets where load is growing fastest and where new supply is hardest to add. Being already connected in a market with a multi-year interconnection queue is worth more than being good at building plants.

Against all three sits one number. Return on invested capital has been 1.0%, 2.3%, 7.0%, 4.9%, negative, negative, 9.5%, 13.8% and 5.6% across the nine years since 20178. Against an 8% cost of capital that is two good years out of nine. Two of the bad ones have honest explanations — Winter Storm Uri in 2021 and hedging marks since — and honest explanations do not pay a dividend.

So: narrow. The assets are real and partly irreplaceable, the integrated structure is a genuine advantage over a pure generator or a pure retailer, and the recent contracting is a serious attempt to convert a commodity business into a contracted one. What has not yet happened is the thing that would make it wide: a demonstrated ability to earn well above the cost of capital across a full cycle rather than in the two years when power prices happened to be high. The trajectory is widening because the twenty-year agreements with Amazon and Meta change the shape of the earnings. Watch the return on capital, not the megawatts.

Moat trajectory: Widening

The assets have not changed and what they are worth has. Twenty-year agreements with Amazon and Meta now cover a little over half the nuclear fleet, converting the highest-margin output in the company from a weather forecast into a schedule. That is a structural change rather than a cyclical one. What holds the direction short of anything stronger is that none of the contracted power has been delivered yet, and the return on capital underneath is still a merchant generator's.

The number that tests this moat
Moat Explorer calc
ROIC vs 8% cost of capital
5.6% in 2025; above 8% twice in nine years

The series since 2017 reads 1.0, 2.3, 7.0, 4.9, negative, negative, 9.5, 13.8, 5.6. Two years above the hurdle out of nine is the honest verdict on merchant generation, and it is why this is rated narrow rather than wide. The thesis is falsified if it does not improve as contracted volumes begin flowing from late 2027.

Source: Vistra Corp. Form 10-K, fiscal year 2025 ↗
Aspects of the moat
DIG DEEPER 01
The Licences Nobody Can Get
Six reactors bought for about $766 a kilowatt — the only fuel in the fleet whose quantity nobody can add to.
DIG DEEPER 02
The Customer at the End of the Wire
$8.5 billion of power sold inside the company — the plants have a customer before the power is made, and it is the same company.
DIG DEEPER 03
Already Connected
The scarce thing in American power is not equipment. It is a place on the grid, and Vistra's 44,000 megawatts are already through the queue.
DIG DEEPER 04
Selling the Output Before It Is Made
Half the nuclear fleet contracted for twenty years to Amazon and Meta — a merchant generator converting itself, reactor by reactor.
References
  1. ReportedVistra owns six nuclear reactors at four sites totalling 6,448 megawatts, with operating licences running to dates between 2036 and 2053
    Vistra Corp. Form 10-K, FY2025, Item 2 Properties and generation fleet — six nuclear generating units at four facilities totalling 6,448 MW: Comanche Peak Unit 1 (ERCOT, 1,200 MW, 18-month refuelling, licence to 2050), Comanche Peak Unit 2 (ERCOT, 1,200 MW, 2053), Beaver Valley Unit 1 (PJM, 939 MW, 2036), Beaver Valley Unit 2 (PJM, 933 MW, 2047), Perry (PJM, 1,268 MW, 24-month refuelling, 2046) and Davis-Besse (PJM, 908 MW, 24-month refuelling, 2037), with nuclear units "generally operated at full capacity"; coal and lignite fleet of seven facilities totalling 8,743 MW; twelve peaking generation facilities totalling 4,822 MW; battery energy storage of 350 MW in California, 270 MW in Texas and 4 MW in Illinois; solar of 538 MW in Texas and 112 MW in Illinois; the Moss Landing 100 MW and 300 MW battery facilities "will not return to service" — FY2025 · publ. February 2026 · source ↗
  2. Moat Explorer calcVistra did not build these either — it bought 4,048 megawatts of them for $3.1 billion, which works out near $766 a kilowatt for plant that is already licensed, already staffed and already connected
    Vistra Corp. Form 10-K, FY2025, growth and portfolio transformation — acquisition of Energy Harbor in 2024 "including 4,048 MW of nuclear generation facilities in PJM"; acquisition in 2025 of 2,557 MW of natural gas generation facilities in Delaware and Pennsylvania (PJM), Rhode Island (ISO-NE), New York (NYISO) and California (CAISO); plans to add 433 MW of uprate capacity from the Perry, Davis-Besse and Beaver Valley nuclear plants in PJM; announced plans "to repower the Coleto Creek and Miami Fort coal generation facilities as natural gas-fueled facilities upon their retirement no later than 2027 and the middle of 2028, respectively"; commercial operations reached at the Oak Hill solar facility in Texas totalling 200 MW; twenty-year PPAs with Meta for a total of 2,609 MW of carbon-free power and capacity from the PJM nuclear plants, with delivery commencing on a portion of operating energy and capacity in late 2026 and full delivery by year end 2027, and uprate delivery commencing in 2031 with full delivery by year end 2034 — FY2025 · publ. February 2026 · source ↗
  3. ReportedFive million customers is not a moat in the switching-cost sense — a Texan can change electricity provider in an afternoon — but it is a structural advantage of a subtler kind
    Vistra Corp. Form 10-K, FY2025, Item 1 Business — "The Company brings its products and services to market in 18 states and the District of Columbia, including all major competitive wholesale power markets in the U.S. We serve approximately 5 million residential, commercial, and industrial retail customers with electricity and natural gas. Our generation fleet totals approximately 44,000 megawatts of generation capacity powered by a diverse portfolio, including natural gas, nuclear, coal, solar, and battery energy storage facilities"; the integrated model "enables us to structure products and contracts in a way that offers significant value compared to stand-alone retail electric providers"; five reportable segments — Retail, Texas, East, West and Sunset, plus Asset Closure; retail investors served through TXU Energy in ERCOT, Homefield Energy in MISO and Public Power in PJM, ISO-NE, NYISO and MISO — FY2025 · publ. February 2026 · source ↗
  4. ReportedVistra's own description is that the integrated model lets it structure products in a way a stand-alone retailer cannot
    Vistra Corp. Form 10-K, FY2025, Item 1 Business — "The Company brings its products and services to market in 18 states and the District of Columbia, including all major competitive wholesale power markets in the U.S. We serve approximately 5 million residential, commercial, and industrial retail customers with electricity and natural gas. Our generation fleet totals approximately 44,000 megawatts of generation capacity powered by a diverse portfolio, including natural gas, nuclear, coal, solar, and battery energy storage facilities"; the integrated model "enables us to structure products and contracts in a way that offers significant value compared to stand-alone retail electric providers"; five reportable segments — Retail, Texas, East, West and Sunset, plus Asset Closure; retail investors served through TXU Energy in ERCOT, Homefield Energy in MISO and Public Power in PJM, ISO-NE, NYISO and MISO — FY2025 · publ. February 2026 · source ↗
  5. ReportedThe $8.5 billion of intersegment sales is what that looks like in the accounts
    Vistra Corp. Form 10-K, FY2025, segment revenue note — total revenues by segment for 2025: Retail $14,340M, Texas $5,353M, East $6,174M, West $325M, Sunset $74M, eliminations $(8,528)M, total $17,738M; intersegment sales Retail $107M, Texas $4,419M, East $4,005M, West $3M, Sunset $(3)M, eliminations $(8,531)M; unrealized hedging revenue $(766)M in 2025 against $1,013M in 2024; realized hedging revenue $583M; transferable PTC revenues $229M; East capacity revenue of $793M sold offset by $566M purchased. For 2024 total revenues were Retail $12,797M, Texas $5,394M, East $5,661M, West $839M, Sunset $39M, eliminations $(7,506)M, total $17,224M; for 2023, Retail $10,572M, Texas $3,979M, East $5,890M, West $866M, Sunset $48M, eliminations $(6,576)M, total $14,779M — FY2025 · publ. February 2026 · source ↗
  6. ReportedVistra's fleet sits in ERCOT, which manages about 83,707 megawatts of Texas peak demand for roughly 27 million customers, and in PJM, which manages about 160,709 megawatts for about 67 million
    Vistra Corp. Form 10-K, FY2025, market discussion — "ERCOT is an ISO that manages the flow of electricity from approximately 83,707 MW of 2025 peak demand to approximately 27 million Texas customers, representing approximately 90% of the state's electric load"; "PJM is an RTO that manages the flow of electricity from approximately 160,709 MW of peak 2025 demand to approximately 67 million customers" across thirteen states and the District of Columbia; "if a less efficient natural gas unit is needed to meet demand, its offer price sets the market clearing price for all dispatched generation in that market, regardless of other units' offer prices"; prices "vary within different zones due to transmission losses and congestion"; under the PUCT-approved Emergency Pricing Program the system-wide offer cap temporarily falls to $2,000/MWh if prices have been at the cap for 12 hours in a rolling 24-hour period, and the maximum point on each ASDC is reduced to $2,000/MWh for the remainder of the calendar year once the Cost of New Entry reference price is exceeded — FY2025 · publ. February 2026 · source ↗
  7. ReportedVistra's fleet sits in ERCOT, which manages about 83,707 megawatts of Texas peak demand for roughly 27 million customers, and in PJM, which manages about 160,709 megawatts for about 67 million
    Vistra Corp. Form 10-K, FY2025, market discussion — "ERCOT is an ISO that manages the flow of electricity from approximately 83,707 MW of 2025 peak demand to approximately 27 million Texas customers, representing approximately 90% of the state's electric load"; "PJM is an RTO that manages the flow of electricity from approximately 160,709 MW of peak 2025 demand to approximately 67 million customers" across thirteen states and the District of Columbia; "if a less efficient natural gas unit is needed to meet demand, its offer price sets the market clearing price for all dispatched generation in that market, regardless of other units' offer prices"; prices "vary within different zones due to transmission losses and congestion"; under the PUCT-approved Emergency Pricing Program the system-wide offer cap temporarily falls to $2,000/MWh if prices have been at the cap for 12 hours in a rolling 24-hour period, and the maximum point on each ASDC is reduced to $2,000/MWh for the remainder of the calendar year once the Cost of New Entry reference price is exceeded — FY2025 · publ. February 2026 · source ↗
  8. Moat Explorer calcReturn on invested capital has been 1.0%, 2.3%, 7.0%, 4.9%, negative, negative, 9.5%, 13.8% and 5.6% across the nine years since 2017
    Moat Explorer calculation from SEC EDGAR XBRL — return on invested capital as NOPAT divided by average operating invested capital, where NOPAT is OperatingIncomeLoss multiplied by one minus the effective tax rate and invested capital is total assets less current liabilities less cash: 1.0% (2017), 2.3% (2018), 7.0% (2019), 4.9% (2020), -5.4% (2021), -4.2% (2022), 9.5% (2023), 13.8% (2024), 5.6% (2025). The 8% hurdle is an assumed weighted average cost of capital, not a filed figure — FY2017-FY2025 · publ. August 2026 · source ↗
Sources
Generated September 23, 2026