Vistra Corp
Owns/operates a multi-tech generation fleet (gas, nuclear, coal, solar, storage) selling power and capacity into wholesale markets (ERCOT, PJM, ISO-NE, NYISO, CAISO), vertically integrated with a large competitive retail arm (TXU Energy, Dynegy, Ambit) that hedges the generation length; increasingly contracting nuclear baseload to hyperscalers via 10-20yr PPAs. Earnings driven by spark spreads, capacity-auction clears and retail margin, not a regulated rate base.
The thesis on this name
State of Data-Center Power
An integrated IPP with nuclear + a large gas fleet, now ~50GW after the Calpine acquisition (~21GW added, fact), and its own Meta nuclear deal. The asymmetry vs Constellation is valuation: Vistra trades ~13x EV/EBITDA against CEG's 17-21x (fact) — a cheaper way to own dispatchable generation into a structurally short grid, with PJM capacity prices hitting the price cap ($329-333/MW-day, +22% YoY, fact) directly lifting its merchant economics. Less of a pure-nuclear-scarcity premium than CEG, more of a 'cheap dispatchable megawatts into record capacity prices' value call.
State of Data-Center Power
The merchant-power + nuclear book leveraged to PJM/ERCOT scarcity pricing and datacenter PPAs — cheaper than CEG, higher beta to the power curve.
State of Data-Center Power
Meta 2.6GW deal + PJM capacity step-up; higher-beta, cheaper sibling to CEG, levered to scarcity pricing.
Earnings, margins, COGS & capex
Vistra's reported revenue is nearly flat (~$17.7B, +3% in FY2025) because it is largely a flow-through of wholesale and retail power prices, while the real earnings engine is generation economics: Ongoing Operations Adjusted EBITDA grew to $5.91B (FY2025) and is guided to $6.8-7.6B in 2026 as the 2024 Energy Harbor nuclear acquisition, record-cap PJM capacity clears, and a recovering retail book compound. Q1 2026 was a record first quarter ($1.49B Adj EBITDA, net income $1.03B), and management reaffirmed full-year guidance that excludes the pending $4.7B Cogentrix gas acquisition and the Meta/AWS nuclear PPAs. The business throws off ~$3.6-4.7B of FCF-before-growth, funding buybacks, a small dividend, and deleveraging toward investment grade (achieved at a second agency in early 2026). Headline GAAP net income is noisy ($944M FY2025) due to ~$0.8B of unrealized hedge mark-to-market.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~0¢ is cost of goods and ~90¢ operating expense, leaving ~10¢ of operating profit (~5¢ net).
Revenue trend
Margins
up
down (noisy)
down then up
up
COGS structure
COGS is dominated by fuel and purchased power: natural gas for the ~30+ GW gas fleet, plus coal/lignite, nuclear fuel, and power purchased to serve retail load when generation is short. Realized/unrealized hedge settlements, ERCOT/PJM ancillary and congestion costs, plant O&M, emissions/environmental compliance, and nuclear decommissioning accruals round it out. Because the model is integrated, much of 'revenue' and 'COGS' net out as a spread, the wholesale fuel-to-power spark/dark spread is the true unit economic.
Capex
Growth+maintenance capex runs ~$2.2-2.6B/yr [estimate], funding: (1) the Vistra Zero buildout (solar + battery storage in Texas/California), (2) nuclear uprates and life-extension capex to deliver the Meta PPA capacity, (3) new dispatchable gas (e.g., a Permian Basin ~860 MW peaker), and (4) maintenance on the existing fleet. The $4.7B Cogentrix and $1.9B Lotus deals are M&A, separate from organic capex.
Latest earnings
Beat: Q1 2026 Adj EBITDA of $1.49B came in above consensus and was a record first quarter; the company swung to a clear GAAP profit and reaffirmed full-year guidance [fact]. (Note: Q4 2025 had been a softer/miss quarter, so Q1 2026 was framed as a rebound.)
Reaffirmed FY2026 Ongoing Ops Adjusted EBITDA $6.8-7.6B and Adjusted FCFbG $3.925-4.725B; guidance EXCLUDES the pending Cogentrix acquisition and the Meta/AWS nuclear PPA contributions, implying upside as those layer in [fact].
- Q1 2026 Adj EBITDA
- $1,494M (record Q1)
- Q1 2026 GAAP net income
- $1,029M
- FY2026 Adj EBITDA guide
- $6.8-7.6B (ex-Cogentrix/PPAs)
- Credit rating
- Investment grade at 2nd major agency (early 2026)
Growth drivers
- Hyperscaler nuclear PPAs — ~3,800 MW contracted (Meta ~2,600 MW PJM nuclear + uprates; AWS up to 1,200 MW at Comanche Peak, 20yr) that re-rate baseload nuclear from spot to long-dated contracted cash flow [fact]
- PJM/ERCOT capacity scarcity — PJM cleared at the FERC cap ($329.17/MW-day 26/27, $333.44 27/28) with the RTO short of its reliability requirement, lifting capacity revenue [fact]
- AI/datacenter load growth tightening reserve margins and lifting forward power and spark spreads in both ERCOT and PJM [fact]
- M&A roll-up of dispatchable gas at ~7x EBITDA — Cogentrix ($4.7B, ~5,500 MW) pending 2H26; Lotus (2,600 MW) closed Nov 2025 [fact]
- Vistra Zero renewables/storage buildout adding contracted clean capacity [fact]
- Deleveraging to investment grade + buybacks compounding per-share value [fact]
Reported financials — SEC EDGAR
Audited GAAP figures pulled from SEC filings · latest filing 2026-02-27. The audited primary-source spine — not financial advice.
Revenue — annual (GAAP)
Margins & balance sheet — FY’25
Bull & bear
VST is the highest-beta, cheapest large-cap way to own a structurally tightening US power curve: a massive dispatchable + nuclear fleet selling into PJM/ERCOT scarcity at record capacity prices, with hyperscaler PPAs converting merchant nuclear into contracted cash flow, all at ~10x EV/EBITDA vs CEG's ~13x, and with Cogentrix + PPA upside not yet in guidance.
- Cheaper than CEG (~10x vs ~13x forward EV/EBITDA) for arguably more torque to the power curve and higher FCF yield, classic relative-value setup for the AI-power trade
- Guidance ($6.8-7.6B 2026 Adj EBITDA) explicitly EXCLUDES Cogentrix (~5,500 MW) and the Meta/AWS nuclear PPAs, so there is visible, contracted upside layering in over 2026-2034
- PJM cleared at the FERC price cap two auctions running with the RTO SHORT of its reliability requirement, this is the tightest capacity backdrop in PJM history and VST is long capacity
- ~$3.6-4.7B/yr FCF funds aggressive buybacks on a shrinking share count plus deleveraging to investment grade (already 2 agencies), compounding per-share value
- Integrated retail + multi-region fleet smooths the commodity cycle vs pure-merchant peers, and nuclear uprates/license renewals extend the best assets 20 years
VST trades at a discount for real reasons: the most leveraged balance sheet among large IPPs, earnings that swing with weather and power prices, coal/ESG overhang, and a stock whose price already embeds an aggressive multi-year AI-load-growth ramp that, if it slows or supply catches up, mean-reverts the spark/capacity prices that drive the whole thesis, and VST falls harder than contracted peers.
- High-beta cuts both ways: a mild-weather, low-volatility year or a datacenter-buildout pause compresses ERCOT scarcity pricing and PJM forward curves, and VST has the least contracted cash flow buffer
- Leverage is the discount: ~$16.2B net debt / ~2.7x means rate or spread shocks and any EBITDA disappointment hit equity disproportionately; deleveraging consumes FCF that bulls want for buybacks
- Capacity upside is administratively capped, PJM already cleared AT the FERC cap, so the 'scarcity re-rate' has a regulator-imposed ceiling, and FERC/state intervention risk is rising as datacenter cost-allocation becomes political
- After a huge multi-year run (52-wk high ~$220), much of the AI-power optimism is priced in; GAAP earnings are noisy (hedge marks) and the stock is volatile (~$133-$220 range)
- Coal/lignite retirement costs and ESG screens, plus hyperscaler counterparty/tech-capex concentration on the long-dated PPAs, are tail risks the multiple doesn't fully reflect
What it is worth
Forward EV/EBITDA relative-value vs IPP peers, cross-checked against FCF yield and guidance-implied EBITDA.
~$110-130 (power-curve/capacity mean-reversion + rate/credit pressure compress to ~8x; revisits the 52-wk-low zone) [estimate]
~$180-200 (re-rate toward ~11-11.5x on guidance midpoint + Cogentrix close + continued buybacks/deleveraging) [estimate]
~$230-260 (12x+ on rising power/capacity curves, full PPA + Cogentrix EBITDA layering in, multiple converges toward CEG) [estimate]
At ~$168 / ~$56.6B mkt cap (~$76B EV), VST trades ~10x 2026E EBITDA (~$7.2B guide midpoint) vs CEG ~13x and recent gas-asset deals at ~7-8x; the discount is leverage + commodity beta, the upside is Cogentrix + Meta/AWS PPAs not yet in guidance.
SWOT
Strengths
- Largest US competitive generation fleet (~41 GW) spanning gas, nuclear, coal, solar and storage, geographically split ERCOT/PJM/ISO-NE/NYISO/CAISO, diversifying weather and basis risk
- Integrated retail (TXU Energy, ~#2 competitive retailer) that internally hedges generation length and stabilizes cash flow vs pure-merchant peers
- Second-largest US competitive nuclear fleet (~6.4 GW) post-Energy Harbor, now anchoring 20yr hyperscaler PPAs (Meta, AWS) at premium prices
- Strong FCF generation (~$3.6-4.7B/yr) funding buybacks, dividend, deleveraging to investment grade, and accretive ~7x EBITDA gas M&A
- High operating beta to a structurally tightening power curve, more upside per dollar than regulated utilities
Weaknesses
- Highest leverage among large IPPs — ~$16.2B net debt, ~2.7x net debt/EBITDA, which is the main reason VST trades at a discount to CEG
- Earnings are commodity- and weather-exposed — GAAP results are noisy (~$0.8B unrealized hedge marks in FY2025) and hard to model quarter-to-quarter
- Coal/lignite exposure carries environmental, retirement-cost and ESG-screen overhang vs nuclear-pure CEG
- Capacity revenue depends on administratively set price caps (PJM cleared AT the cap), creating regulatory/political ceiling risk on the upside
- Capital-intensive: large maintenance + uprate + growth capex needed just to deliver contracted PPA capacity
Opportunities
- Datacenter/AI load is driving multi-year PJM/ERCOT scarcity — each cleared capacity auction and forward-curve uplift is high-incremental-margin
- More hyperscaler PPAs (behind-the-meter and grid-connected) on its nuclear and new gas, converting merchant length to contracted cash flow
- Cogentrix (~5,500 MW gas) and further dispatchable-gas roll-ups at ~7x EBITDA when peers (NRG, CEG) are also acquiring, scale advantage
- Nuclear uprates + subsequent license renewals (20yr extensions on all four PJM units) extend the highest-margin asset's life
- Vistra Zero solar/storage growth + potential SMR/co-located generation optionality
Threats
- Power-price mean reversion — if AI load growth disappoints or new supply (gas, solar+storage) floods in, forward spark spreads and capacity prices compress, hitting VST harder than contracted peers
- FERC/state intervention capping capacity prices or re-allocating datacenter interconnection costs (PJM already at administrative caps)
- Mild weather / low-volatility years depress ERCOT scarcity-pricing upside (the high-beta thesis cuts both ways)
- Rising rates / credit-spread widening raise the cost of its sizable debt load and slow deleveraging
- Hyperscaler counterparty concentration and tech-capex pullback risk on the long-dated PPAs; nuclear operational/outage risk
Moats, dependencies & bottlenecks
Moats
Irreplaceable nuclear + dispatchable baseload assets (2nd-largest US competitive nuclear fleet, ~6.4 GW; ~41 GW total) New nuclear is effectively un-buildable on this timeline; existing units with 20yr license renewals are scarce, carbon-free, firm capacity hyperscalers are paying premiums for [fact]
Scale + integration (largest competitive fleet + #2 competitive retail book) Generation-plus-retail integration internalizes the hedge and gives cost/dispatch advantages; scale enables ~7x-EBITDA roll-ups peers can't all match
Locational/grid interconnection in the right markets (ERCOT + PJM scarcity zones near datacenter demand) Sited generation with firm interconnection near load is scarce given multi-year interconnection queues, but markets/regulation can shift value
20yr take-or-pay-style contracts convert merchant exposure to contracted cash flow, but it is contract-by-contract, not a structural barrier to entry [fact]
Investment-grade balance sheet + FCF scale enabling accretive M&A IG access and ~$4B+ FCF let VST acquire dispatchable gas (Cogentrix, Lotus) cheaply, but leverage is higher than CEG so the edge is conditional
Dependencies
The ~30+ GW gas fleet's economics and retail-supply cost hinge on gas; gas is both fuel COGS and the marginal price-setter, a double-edged sensitivity
PPA value depends on tech-sector capex durability and counterparty credit; concentration risk if AI-datacenter spend slows [fact]
Capacity revenue is set by auction rules and price caps (cleared at the FERC cap); datacenter cost-allocation and price-collar decisions directly move EBITDA [fact]
ERCOT scarcity pricing and the high-beta thesis depend on hot summers / volatile load; mild years compress the upside
Delivering the Meta/AWS PPA capacity requires on-time uprates (2031-2034) and license renewals; outage or execution slippage delays contracted cash flow [fact]
Advantages
- Highest torque to the power curve among large caps, more EBITDA upside per dollar of rising spark/capacity price than regulated utilities or contracted CEG
- Cheapest large IPP (~10x EV/EBITDA vs CEG ~13x) with higher FCF yield, relative-value entry into the same AI-power theme
- ~41 GW multi-tech, multi-region fleet diversifies weather/basis risk while retaining merchant upside
- ~$3.6-4.7B/yr FCF self-funds buybacks, dividend, deleveraging AND ~7x-EBITDA accretive gas M&A simultaneously
- Scarce carbon-free firm nuclear that hyperscalers are competing to contract, a seller's market for exactly VST's best assets
- Investment-grade balance sheet (2 agencies) lowers cost of capital and widens the M&A funnel
Weaknesses
- Most leveraged large IPP (~$16.2B net debt, ~2.7x), the structural reason for its valuation discount and its rate-sensitivity
- Earnings are commodity/weather-driven and GAAP-noisy (~$0.8B unrealized hedge marks in FY2025), low visibility quarter-to-quarter
- Coal/lignite fleet carries retirement-cost, environmental and ESG-screen overhang absent at nuclear-pure CEG
- Capacity-price upside is administratively capped and politically exposed (datacenter cost-allocation fights)
- High stock volatility (~$133-$220 52-wk range) and a multi-year run mean a lot of AI-power optimism is already priced in
- Counterparty + execution concentration on the long-dated nuclear PPAs (hyperscaler credit + uprate delivery 2031-2034)
Bottlenecks
- Multi-year grid interconnection queues in PJM/ERCOT limit how fast new generation (and new datacenter load) can connect, capping organic growth speed
- Nuclear uprate engineering, NRC licensing and outage windows gate when contracted PPA megawatts actually deliver cash (Meta uprates phased 2031-2034)
- Leverage/credit ceiling — ~2.7x net debt/EBITDA constrains how much M&A and buyback can run simultaneously without straining the IG rating
- FERC-imposed capacity price caps cap the per-MW upside even when the market would clear far higher (PJM shadow clear ~$141,828/MW-yr vs capped result)
- Skilled-labor, turbine and transformer supply-chain lead times for new dispatchable gas and storage buildout
Top signals & trends
Top signals
The core driver, sustained tightness + reliability shortfalls keep capacity revenue at the cap; watch for cap changes or supply response [fact]
Each is guidance-additive since current 2026 guide excludes both; a closed Cogentrix re-rates the EBITDA base [fact]
Deleveraging + IG ratings are the path to closing the valuation discount vs CEG [fact]
The whole thesis rests on this ramp continuing; a slowdown is the single biggest reversal risk [fact]
Hot, volatile summers drive scarcity upside; market-design changes alter how ancillary/energy value accrues
Regulatory intervention is the asymmetric downside to the capped-capacity upside
Trends
Structural multi-year scarcity, PJM 27/28 peak forecast +5,250 MW, ~5,100 MW of it datacenters; RTO short of reliability requirement [fact]
Premium long-dated prices for carbon-free firm baseload re-rate VST's nuclear earnings quality [fact]
Vistra (Cogentrix, Lotus), NRG (LS Power 13 GW), CEG (Calpine), scale + cheap M&A favors the largest balance sheets [fact]
Caps both protect against demand-destruction backlash and ceiling the upside; cost-allocation fights add political risk [fact]
Retires VST's lowest-quality assets and tightens supply (positive), but raises retirement-cost and intermittency-management burden
Pressures VST's sizable debt load and the valuation of long-duration contracted cash flows more than for lower-levered peers
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Permian gas, Energy Transfer, Kinder Morgan) Primary fuel for the ~30+ GW gas fleet; gas basis and pipeline capacity drive COGS
e.g., Cameco, Centrus) Fuel for the ~6.4 GW nuclear fleet; supply security matters for the long-dated PPAs
Siemens Energy, Mitsubishi Power) Gas turbines, uprate hardware, and grid equipment with multi-year lead times
Battery/solar suppliers for Vistra Zero (e.g., Tesla, Fluence, panel makers) Storage and solar inputs for the renewables buildout
owned Texas lignite mines) Fuel for the remaining coal fleet; declining as units retire
20yr PPAs for >2,600 MW from PJM nuclear plus uprates (2031-2034), anchor AI-load customer [fact]
20yr PPA up to 1,200 MW carbon-free at Comanche Peak (delivery from late 2027) [fact]
Dynegy, Ambit, Value Power) ~#2 US competitive retailer; ~5M+ residential/commercial accounts, added ~1M in PJM in 2024 [fact]
ERCOT / PJM / ISO-NE / NYISO / CAISO wholesale markets Energy + capacity + ancillary-services buyers of merchant output
Commercial & industrial / datacenter offtakers Pipeline of additional hyperscaler and C&I PPAs beyond Meta/AWS
The direct comp and premium peer, largest US nuclear fleet (~60 GW total), trades ~13x EV/EBITDA vs VST ~10x; the board's framing is VST = cheaper, higher-beta CEG [fact]
Integrated retail + generation competitor, acquiring 13 GW of LS Power gas at ~7.5x; overlaps in ERCOT retail and gas [fact]
Merchant + nuclear (Susquehanna) IPP, a pure-play high-beta peer that pioneered behind-the-meter datacenter deals (Amazon) [fact]
Largest US gas+geothermal fleet; CEG's acquisition of Calpine creates a gas+nuclear giant directly overlapping VST's gas book [fact]
PJM nuclear (regulated + merchant); a contracted-nuclear alternative for the same hyperscaler-PPA demand
Regulated peers compete for datacenter load via rate-based generation; lower-beta substitute for the same AI-power theme