Vistra Corp. (NYSE: VST) — Best Operator in a Commodity Business, Re-Priced From Euphoria to Fair

Report date: 2026-06-13 Price (2026-06-12): $148.02 · 52-wk range: $132.66 – $219.82 · Market cap: ~$50B · Enterprise value: ~$71.7B · Shares out: ~337–338M Sector / sub-industry: Utilities — Independent Power Producers (merchant generation + competitive retail)

⚡ Claude’s Take

This block is the author’s own independent opinion and general information only — not investment advice. Everything below it — the full analysis that follows — carries no recommendation and no price target; only this block takes a position.

Verdict: HOLD / accumulate-on-weakness in the low-$120s to low-$130s. Not a short; not a back-up-the-truck buy at $148. Vistra is the fossil-heavy, lower-quality, cheaper mirror of Constellation (CEG): a genuinely best-in-class capital allocator — it retired ~28% of its share count at an average cost under $36 while the stock now trades near $150 — running a structurally bad, no-moat commodity business (a MWh is the canonical commodity) that happens to be enjoying a once-in-a-generation AI-power demand window. After a ~33% drawdown from a $219.82 high struck in the 2024–25 AI-utility euphoria, the stock is no longer priced for a fantasy. But it is not cheap, either — at ~10x forward EV/EBITDA (ex-Cogentrix) and ~8.6% FCF-before-growth yield it sits at the 77th percentile of its own ten-year valuation history (P/S 93rd, P/B 86th). The de-rate removed the bull premium and re-priced VST back to roughly its own base case — fair value, not deep value.

What separates VST from CEG, and caps my enthusiasm, is the asset mix and the geography. Only ~15% of capacity is nuclear (versus CEG’s ~two-thirds), so the IRA Section 45U production-tax-credit floor — the one hard government backstop in this industry — protects a far smaller slug; the other ~82% (gas + retiring coal) is pure merchant price-taking. And ~46% of the fleet sits in ERCOT, an energy-only market with no capacity payments, post-Uri price caps that truncate the upside, and a supply response that is already arriving — batteries doubled to ~16 GW, Texas-Energy-Fund-subsidized gas is interconnecting, and management itself confirms ERCOT forwards are softening and that batteries have “returned virtually nothing.” That is the Marathon capital-cycle mean-reversion mechanism biting in VST’s single largest market, now. The offsets are real: a ~5M-customer retail book that is a genuine counter-cyclical hedge (not a moat), ~3.8 GW of nuclear newly contracted under 20-year AWS/Meta PPAs (more signed nuclear than any U.S. peer), an investment-grade balance sheet (S&P BBB- since Dec-2025), and the relentless buyback. Framing: a high-quality operator and capital allocator trapped in a structurally cyclical, no-moat industry, re-priced from euphoria to fair — buy the cash-return engine on weakness, not the AI narrative at full price. I’d be a committed buyer in the low-$120s/$130s where the ~9–10% FCFbG yield and the buyback do the heavy lifting; a patient holder here; and not a seller, because the cash generation and sub-$36-cost-basis buyback put a real floor under the equity.

Conviction: medium. Flips bullish if ERCOT forwards stabilize or firm (the supply flood absorbed by demand rather than crushing margins) and VST signs additional nuclear/gas data-center PPAs at disclosed premiums while Cogentrix proves clearly accretive — that converts merchant exposure to contracted annuity and justifies holding the re-rated multiple. Flips bearish if ERCOT power forwards structurally re-base lower as batteries/TEF-gas/solar overwhelm the demand ramp, the PJM capacity windfall is clawed back, and the AI-era merchant multiple mean-reverts toward its pre-2023 high-single digits — that combination strips both EBITDA and the multiple at once and is the genuine path to the ~$100 bear zone. Tag: cheaper than Constellation for a reason — and Texas is already supplying its way back toward the mean.

1. Executive Summary

Vistra Corp. is the largest competitive (merchant) power producer in the United States by generation volume and the largest competitive electricity retailer, serving ~5 million customers across 18 states and the District of Columbia from a fleet of ~43,641 MW (pre-Cogentrix) of natural gas, nuclear, coal, solar, and battery capacity. It was forged out of the 2016 bankruptcy reorganization of Energy Future Holdings (the old TXU), bulked up via the 2018 Dynegy merger, the 2019 Ambit retail deal, the March-2024 Energy Harbor nuclear acquisition, the 2025 Lotus gas deal, and the pending (2H-2026) ~$1.9B Cogentrix gas acquisition. It is run by CEO Jim Burke and CFO Kris Moldovan from Irving, Texas.

The investment tension is the same one that governs the entire merchant-power complex, sharpened by VST’s specific asset mix. Merchant generation is, historically, a structurally bad business: a megawatt-hour is a perfect commodity, price is set by the marginal unit, there are no barriers to entry, and the sector’s capital cycles have repeatedly bankrupted their participants (Calpine, NRG, Mirant, Dynegy, and VST’s own TXU predecessor). Greenwald’s framework is unforgiving here — no supply, demand, or scale advantage means ROIC mean-reverts to the cost of capital. Against that prior, two things distinguish VST: (1) a ~6.4 GW nuclear fleet that carries the same federal 45U PTC revenue floor as CEG’s — but at ~15% of capacity, not two-thirds; and (2) a demonstrated operating and capital-allocation competency — buying assets cheaply in distress, integrating them, and routing the cash into a share count that has shrunk ~28% in four years. The first is a narrow, government-granted moat on a minority of the business; the second is a real edge but a capability, not a barrier to entry.

The de-rate is the story. From a ~$219.82 peak set during the AI-power mania, VST has fallen ~33% to ~$148, sitting near its 52-week low. Unlike CEG — which fell ~40% but stayed at a clear premium to the merchant pack — VST’s drop has re-priced it to a mid-pack merchant valuation: ~10x forward EV/EBITDA and ~13.5x a Street-constructed forward P/E, cheaper than CEG (~14x / ~18.6x) and the regulated comps, but dearer than NRG (~6.9x). Crucially, the drawdown compressed the multiple off a bubble peak rather than down to a distressed own-history low: VST remains at the 77th percentile of its own ten-year valuation. The bull must therefore argue the AI-era re-rating is permanent, not that VST is cheap versus itself — because on its own history, it is not.

What the analysis finds: (1) No durable franchise moat — a commodity merchant plus a competed, low-switching-cost retail book, overlaid by a 45U floor on a minority nuclear slug and a genuine acquire-integrate competency; “best operator in a bad industry.” (2) GAAP earnings are noise — non-cash mark-to-market on commodity hedges swung net income from +$2,659M (FY24) to +$944M (FY25) while Adjusted EBITDA rose; value VST only on Adjusted EBITDA and Free Cash Flow before Growth (FCFbG), never P/E or P/B. (3) A genuine cash machine — Adjusted EBITDA roughly tripled ($1.9B FY21 → $5.84B FY25), FCFbG grew to ~$3.6B (guided $3.9–4.7B FY26), funding the buyback and an IG upgrade — but the cash is commodity-/weather-exposed, lumpy (collateral swings turned CFO negative in 2021–22), and excludes a large, back-end-loaded growth-capex program. (4) A structurally cyclical industry at or near a cyclical high — PJM capacity at the auction cap (down only via clawback), ERCOT forwards softening as supply arrives — where VST’s geographic mix concentrates its earnings in the most-contested, least-protected corner (energy-only ERCOT). (5) Strong-but-tested capital allocation — a value-creating buyback and a well-aligned, per-share-based incentive plan, now plowing rising debt and dilutive equity into gas-asset M&A at a euphoric moment, with insiders selling (all 10b5-1) and none buying the dip.

No recommendation and no price target appear below this summary; the body discusses valuation only as embedded expectations and scenarios. The single deliberate exception is the labeled Claude’s Take block above.

2. Business Overview

What Vistra is. Vistra is an “integrated retail electricity and power generation company” — a merchant power producer bolted to the largest competitive retail electricity book in the United States. It owns power plants, sells the output into wholesale markets (ERCOT, PJM, ISO-NE, MISO, NYISO, CAISO) or under bilateral contracts, hedges that output forward, and separately serves ~5 million end customers through a portfolio of retail brands. It is not a regulated utility: it owns no rate base, earns no allowed return, and bears commodity-price risk. (FACT — 10-K FY2025, Item 1.)

The integrated model — the core structural argument. Management’s central claim is that generation and retail form a natural hedge: generation is a short physical position (you must sell power), retail load is a long demand position (you must buy power to serve customers), so the two offset and the combined entity carries lower earnings volatility than either standalone. Q1-2026 was a live demonstration — an exceptionally mild ERCOT winter hurt the retail segment (lower demand, compressed margins), but the generation segment more than offset it, producing a record calendar-first-quarter Adjusted EBITDA of ~$1.49B. (FACT — Q1-2026 call, 2026-05-07.) This is genuine diversification and a real volatility damper. Whether it is a moat is a separate question, addressed below (it is not).

Reporting structure. Through FY2025 VST reported five segments: Retail, Texas (ERCOT generation), East (PJM/ISO-NE/MISO/NYISO generation, now including the Energy Harbor nuclear), West (CAISO), and Asset Closure (retiring/decommissioning coal and the Moss Landing battery remediation). FY2025 segment Adjusted EBITDA: Retail $1,622M, Texas $1,834M, East $2,282M, West $244M, Asset Closure –$74M, Corp –$70M. Management has begun guiding on a simplified two-bucket “Ongoing Operations” basis — Generation (~$4,290M FY25) and Retail (~$1,622M FY25) — which strips the perennially-negative Asset Closure segment and flatters the reported total by ~$75–150M/year. (FACT — 10-K FY2025 segment tables; Q3/Q4-2025 calls.)

The fleet (12/31/2025, pre-Cogentrix): 43,641 MW. Natural gas 26,989 MW (62%) — 28 combined-cycle plants (22,167 MW) plus 12 peakers; coal 8,743 MW (20%), being retired/converted; nuclear 6,448 MW (15%) — six units at four sites; solar/battery 1,274 MW (3%). By region: Texas/ERCOT 19,858 MW (46%), East 22,254 MW (51%), West 1,529 MW (3%). Expected annual generation exceeds 230 TWh. Fleet utilization runs ~60%, leaving substantial headroom to absorb load growth without new build. (FACT — 10-K FY2025, Items 1–2.)

The nuclear fleet — six units, four sites. Comanche Peak Units 1 and 2 (2,400 MW combined, ERCOT, licensed to 2050/2053) are the legacy Texas plant; Beaver Valley Units 1 and 2 (1,872 MW, PJM), Perry (1,268 MW, PJM), and Davis-Besse (908 MW, PJM) came via the March-2024 Energy Harbor acquisition. The nuclear fleet is the only part of VST with a government-backed revenue floor (the 45U PTC; see the Industry section). (FACT — 10-K FY2025.)

The retail book — ~5 million customers. TXU Energy (a 20±year trademarked brand, ~2.6M Texas customers) anchors a portfolio that also includes Ambit, Dynegy Energy Services, Energy Harbor retail, Homefield, TriEagle, Public Power, and US Gas & Electric. Retail is capital-light and genuinely high-return, and it is the counter-cyclical hedge described above. FY2025 retail Adjusted EBITDA of $1,622M was a record, flattered by non-repeating Energy Harbor integration gains; management explicitly guides to a normalized run-rate of ~$1.4B. (FACT — 10-K; Q4-2025/Q1-2026 calls.)

How it makes money — the revenue stack. (1) Merchant energy — wholesale and retail sales of generation at market prices; the commodity-exposed core. (2) Capacity payments — only in PJM/ISO-NE/NYISO (the East segment); ERCOT is energy-only and pays no capacity. (3) Ancillary services and ERCOT scarcity pricing — paid when the grid is tight. (4) Nuclear 45U PTC — the downside floor on ~6.4 GW. (5) Retail margin — unit margins on ~5M customers’ load. The mix matters enormously: VST’s single largest geography (ERCOT) earns no capacity annuity, only volatile scarcity-energy margins, which is the heart of the bear case.

Verdict (business model): A merchant generation + competitive-retail business whose distinguishing features are (a) the largest competitive retail book in the U.S., providing a real natural hedge that lowers earnings volatility, and (b) a ~6.4 GW nuclear slug with a federal PTC floor. It is not a regulated annuity; it is a commodity merchant with a partial hedge and a partial government backstop, run by an unusually good capital allocator. Cash-flow generation is real but commodity-, weather-, and forward-curve-dependent.

3. Industry Dynamics

The base rate is bad. Through both the Greenwald and Marathon lenses, merchant competitive power generation is a structurally bad industry. Output is a perfect commodity, there is no differentiation, price is set by the marginal generating unit, and the sector has run repeated boom/bust capital cycles that destroyed equity (the early-2000s merchant build-out bankrupted Calpine, NRG, Mirant, NEG; VST’s own ancestor TXU/Energy Future Holdings was the largest leveraged-buyout bankruptcy in history). In Greenwald’s terms: no barriers to entry → strategy is irrelevant, only operating efficiency matters, and ROIC mean-reverts to WACC. Any bull case must specify what, precisely, overrides that prior — and for how long.

The defining structural fact: VST is not CEG. Constellation Energy (CEG), the other large merchant generator, offers a useful contrast; the critical distinction for VST is geographic and fuel mix. VST’s fleet is ~46% ERCOT (energy-only, no capacity market), ~51% PJM-led East (capacity market), ~3% CAISO — and only ~15% nuclear. CEG is ~two-thirds nuclear with a federal PTC floor under most of its output. VST therefore sits on the less-protected side of the same demand window.

ERCOT — energy-only, capped on the upside, and supplying its way back toward the mean. ERCOT is unique among U.S. markets in having no capacity payments: generators are compensated only through energy and ancillary-service prices, supplemented by scarcity-pricing adders when reserves run thin. This rewards VST’s flexible gas and batteries only when scarcity actually shows up — higher-volatility, lower-floor than CEG’s PJM capacity annuity. Three post-Uri (2021) features now truncate the upside: the Operating Reserve Demand Curve was replaced by a redesigned scarcity adder (effective Dec-2025); a “peaker net margin” circuit-breaker and the PUCT Emergency Pricing Program both chop the effective system cap toward ~$2,000–4,500/MWh (down from the old $9,000/MWh that made 2021–2023 so lucrative); and the proposed Performance Credit Mechanism — a quasi-capacity construct that would have added ~$1B/year of generator revenue — was shelved by the PUCT. ERCOT stays energy-only; no capacity floor is coming. Meanwhile the Texas Energy Fund is actively subsidizing new gas entry ($5B+ of 3%-rate state loans; first project interconnected April 2026) — VST’s own 10-K flags the TEF as a risk that “may materially change market fundamentals,” i.e., depress incumbent energy prices. (FACT — 10-K FY2025 Item 1A; PUCT; web, 2026.)

The ERCOT supply response is already arriving — the Marathon tell. Battery storage in ERCOT roughly doubled to ~16 GW by mid-2026, solar continues to surge, and TEF-subsidized gas is interconnecting. Management itself confirms this is softening ERCOT forward power prices and states bluntly that ~three years of battery builds have “returned virtually nothing.” This is classic late-boom mean reversion — capital flooding a hot theme, depressing the very margins that attracted it — and it is biting in VST’s single largest market right now. (FACT — Q4-2025/Q1-2026 calls; ERCOT data, web 2026.)

PJM (~51%) — the CEG-style windfall, politically contested. The PJM 2026/27 and 2027/28 Base Residual Auctions both cleared at the administrative price cap (~$329–333/MW-day; they would have cleared materially higher uncapped), a near-term tailwind to VST’s East segment. But this windfall is the explicit target of a political clawback (governors, FERC), faces a one-time September-2026 Reliability Backstop Auction, and is entangled with the December-2025 FERC colocation rulings (relevant to data-center deals). VST’s exposure here mirrors CEG’s, but VST has far less of the irreplaceable, PTC-floored nuclear scarcity underneath it.

Demand — real, but separate the hype from the interconnection reality. ERCOT’s headline large-load interconnection queue ballooned to ~226 GW (+270% YoY) — a number management dismisses as a “low bar” producing “unrealistic projections.” VST’s own underwriting is far more conservative: ERCOT load +5–6%/year through 2030 (peak +3–5%), implying only ~30–40 GW of total ERCOT growth, of which ~10–15 GW is large data centers; PJM +2–3%/year. That is robust versus the ~1–2% historical norm and a genuine tailwind — but a fraction of the queue hype, and management argues much of it can be met by raising utilization of the existing ~60%-utilized fleet rather than new build (bullish for the installed base, bearish for a new-build supercycle). The conservatism is a credibility positive that also caps the blue-sky. (FACT — Q4-2025/Q1-2026 calls.)

Policy. The IRA Section 45U nuclear PTC (up to $15/MWh, phasing out as gross receipts rise between ~$26.00 and ~$44.75/MWh in 2025, inflation-indexed through 2032) is the only hard floor — and it covers just ~6.4 GW of ~44 GW. VST recognized $545M (FY24) then $220M (FY25) of transferable nuclear PTC; the drop as power prices rose confirms it is a downside floor, not a driver. The 2025 “One Big Beautiful Bill” (OBBBA) accelerates the solar/wind credit phase-out — a material headwind to Vistra Zero (VST has confirmed deferral/abandonment of some solar/battery capex) but also a partial offset (it slows the renewable flood depressing ERCOT prices). Federal coal/gas environmental pressure is receding (proposed EPA GHG-rule repeal). (FACT — 10-K FY2025; web 2026.)

Cyclicality. Power prices are closer to a cyclical high than a low: PJM capacity at the cap (downside only), ERCOT forwards softening, post-Uri caps on the right tail. 2026–2027 is heavily hedged (near-term locked), so the cyclical risk is concentrated in 2028+, as hedges roll into a softer, more-supplied curve. VST is implicitly long natural gas (gas sets the marginal price in ERCOT), with the integrated retail book as the counter-cyclical hedge.

Verdict (industry): structurally cyclical, no-moat commodity industry in a favorable-but-maturing demand phase — and structurally weaker for VST than for CEG. Merchant power has zero entry barriers and mean-reverting returns. VST sits on the less-favorable side of the AI window: ~46% of its fleet is in energy-only ERCOT with no capacity floor, truncated upside, and a fast-arriving subsidized + battery supply response already softening forwards; its ~51% PJM exposure rides the same politically-contested capacity windfall CEG depends on but with far less PTC-floored nuclear. Capital-cycle location: late-boom / early supply response, with ERCOT meaningfully closer to the top of the cycle than PJM. The offsets (retail hedge, firm-gas scarcity from the turbine bottleneck, a genuine ~5–6%/year demand tailwind) are real but bounded.

4. Competitive Position

The moat question, asked precisely. A moat exists only if removing it would cause returns to deteriorate toward the cost of capital. For VST, we test each candidate advantage against Greenwald’s taxonomy (supply/cost, demand/captivity, economies-of-scale + captivity) and the financial outcome it should produce.

Generation: no moat. Eighty-two percent of capacity is gas and coal — mid-merit-order thermal that is a pure price-taker on power and spark spreads. There is no supply advantage (fuel and technology are available to all), no demand captivity (the grid buys the cheapest cleared MWh), and no scale-plus-captivity (the market clears on marginal cost regardless of fleet size). This is the canonical commodity with zero barriers to entry. The genuinely new entrants — TEF-subsidized gas, merchant batteries — are arriving precisely because returns looked attractive, the Marathon signature.

Nuclear: a narrow, government-granted edge — but small. The ~6.4 GW nuclear fleet carries the same two advantages CEG’s does: (a) a near-absolute regulatory barrier to new nuclear (irreplaceable, supply-starved carbon-free baseload), and (b) the 45U PTC revenue floor. This is a genuine, financializable advantage — but at ~15% of capacity it protects a far smaller share of VST than of CEG (~two-thirds). The irreplaceable-asset moat is real and proportionally minor.

Retail: demand captivity, but weak and competed. TXU is a real brand with ~20 years of equity and ~2.6M Texas customers; the broader book serves ~5M. Greenwald would classify this as weak demand-side captivity from habit, brand, and switching inertia — in a market (ERCOT’s Power-to-Choose) deliberately engineered for easy switching. Retail is high-ROIC and a real earnings stabilizer, but it is a competed book with low switching costs, not a structural barrier. The share-stability test is suggestive but not decisive: VST’s ERCOT residential share (~32% post-Ambit) leads but is contested by NRG and is stable-to-eroding at the margin.

The one genuine, financializable edge is capability, not a barrier: capital allocation + acquire-integrate-at-scale. VST has repeatedly bought assets cheaply in distress (Dynegy out of the post-2008 merchant bust, Energy Harbor out of FirstEnergy’s bankruptcy reorganization, Lotus, Cogentrix), integrated them, and routed the cash into a share count that fell ~28% in four years at an average cost under $36. That is a real, demonstrated cost-discipline/operating competency that has compounded per-share FCF — but it is a management capability, not an entry barrier, and it is precisely the kind of edge that erodes if the team changes or overpays late in the cycle.

The headline ROE is a trap. VST’s ~43% trailing ROE is not a moat signal: it is flattered by (a) a thin, post-bankruptcy, buyback-shrunk equity base (~$5–5.6B equity on ~$17–19B debt; D/E ~3x) and (b) cyclically peak power prices. Through-cycle, with mean-reverting merchant margins and a normalized equity base, returns are far more pedestrian. Do not read 43% ROE as evidence of franchise quality.

Direct comparison. On the quality axis (nuclear scarcity + contracting), the order is CEG > VST ≈ Talen > NRG. On the cheapness axis, TLN < NRG < VST < PEG < CEG. VST is the mid-quality / mid-price merchant: you pay less than CEG and accept a more commodity-exposed, less-moated, more-ERCOT-concentrated book, backed by a stronger capital-return engine than most peers.

The AI/data-center thesis — genuinely contracted, but read it carefully. VST has signed ~3.8 GW of nuclear under 20-year PPAs — more than any U.S. power company: a 1,200 MW AWS deal at Comanche Peak (Amazon sites a facility on VST land, brings 1-for-1 backup gen; energization Q4-2027, full ramp Q4-2032; options on uprates/SMRs), and January-2026 Meta PPAs covering 2,176 MW of operating capacity (Perry + Davis-Besse) plus 433 MW of uprates. The Meta deals are front-of-the-meter (not colocation-dependent), insulating them from PJM colocation-rule risk. Management projects ~25% adjusted-FCFbG accretion at full nuclear-PPA ramp. The skeptic’s distinction: the operating-capacity PPAs largely re-price existing output — the plant keeps flowing to the grid, but the price is now contracted rather than merchant. That is valuable de-risking (converting commodity exposure to a 20-year annuity), but it is not incremental MW; the genuinely new capacity is only the ~633 MW of long-dated uprates (2031–2034). VST is also steering toward gas data-center structures (colocation, bridge power, new-build gas with the customer taking gas risk), which are broader but lower-premium than CEG/Talen’s pure carbon-free-baseload scarcity play.

Verdict (competitive position): no durable franchise moat — a crowded commodity market with weak differentiation, overlaid by a 45U floor on a minority nuclear slug, a weak/competed retail captivity, and a strong capital-allocation capability. “Best operator in a bad industry” is the honest characterization. The only advantages that survive the financial-outcome test are the narrow nuclear scarcity/PTC floor (small) and the demonstrated capital-allocation competency (real but a capability, not a barrier). The integrated gen+retail model is diversification, not a moat. The ~3.8 GW of signed nuclear PPAs is the single best new development — it converts merchant exposure to contracted annuity on the highest-quality part of the fleet.

5. Growth History and Forward Opportunities

Reported revenue is the wrong lens. Revenue grew from $12,077M (FY21) to $17,738M (FY25) — but this figure is gross-settled, riddled with commodity pass-through and large intersegment eliminations (FY25 elimination/corporate –$8,528M), and rose only ~47% while Adjusted EBITDA tripled. EV/Revenue is meaningless for VST. (FACT — 10-K FY2025.)

Adjusted EBITDA is the real growth story — and it is largely acquired and cyclical, not organic. Total segment Adjusted EBITDA: $1,908M (FY21, depressed by Winter Storm Uri’s –$236M Texas hit) → $2,994M (FY22) → $4,101M (FY23) → $5,539M (FY24) → $5,838M (FY25). That ~tripling decomposes into three drivers: (1) the post-Uri cyclical recovery of ERCOT/PJM power prices off a 2021 trough; (2) the Energy Harbor acquisition (closed March-2024), which added ~$2B+ of East-segment EBITDA on a full-year basis; and (3) genuine operating execution. Very little is organic volume growth — generation volume is roughly flat (the fleet is ~60% utilized, not capacity-constrained). This is high-quality in the sense that it converts to cash and funded buybacks, but it is not a secular volume compounder; it is a cyclical-price + acquisition story. (FACT — 10-K segment tables.)

FY2025 was flattered and partly won’t repeat. Retail Adjusted EBITDA of $1,622M (a record) leaned on non-repeating Energy Harbor integration/supply gains; management guides retail down to a ~$1.4B run-rate. Generation benefited from still-elevated power prices. The honest read: FY2025 is near a cyclical-and-integration high for the legacy business, with the next leg of growth dependent on Cogentrix, the PPA ramp, and demand absorbing the ERCOT supply flood.

Forward opportunities, ranked by quality:

1. Contracted nuclear PPAs (highest quality) — the ~3.8 GW of AWS/Meta deals ramping 2027–2032, projected at ~25% FCFbG accretion at full ramp. This is the genuine de-risking growth.
2. Cogentrix integration (~5.5 GW gas, closes 2H-2026) — adds EBITDA (excluded from current guidance), bought at ~$730/kW (cheap vs new-build), but at a cyclically elevated moment for gas-asset values.
3. Nuclear uprates (~633 MW, 2031–2034) and Permian gas — incremental firm capacity into a tight firm-power market.
4. ERCOT/PJM demand absorption — if the ~5–6%/year load growth materializes and supply does not overwhelm it, merchant margins hold; this is the swing factor and the lowest-certainty lever.
5. Vistra Zero solar/battery — now impaired by OBBBA credit phase-outs and the fact that ERCOT batteries have “returned virtually nothing”; management is deferring/abandoning capex here. A diminished opportunity.

Verdict (growth): mixed quality — cyclical-price + acquisition-driven historically, with the highest-quality forward growth (contracted nuclear PPAs) genuine but modest, and the lowest-quality (merchant margin in oversupplied ERCOT) the largest swing. Do not extrapolate the FY21→FY25 EBITDA tripling; much of it was a price recovery off a trough plus a one-time acquisition step-up. The per-share growth engine has been the denominator (buybacks), not organic EBITDA volume.

6. Financial Quality

Rule one: discard GAAP earnings. VST’s GAAP net income is dominated by non-cash, unrealized mark-to-market on its commodity hedge book and is genuinely uninformative for valuation. GAAP net income: –$1,274M (FY21), –$1,227M (FY22), +$1,493M (FY23), +$2,659M (FY24), +$944M (FY25); GAAP diluted EPS gyrated $3.58 → $7.00 → $2.18 (FY23–25). FY25 net income fell $1,868M year-over-year while Adjusted EBITDA rose $299M, driven by a –$1,963M unfavorable swing in unrealized hedge MTM with zero cash impact. The trailing P/E of ~24.7x is meaningless; even the “13.5x forward P/E” is a Street MTM-stripped construct, usable only directionally. (FACT — 10-K FY2025 MD&A.)

Adjusted EBITDA — the metric that matters. Roughly tripled FY21→FY25 (see the Growth section). FY2026 guidance (introduced Q3-2025, reaffirmed through Q1-2026, all ex-Cogentrix): Adjusted EBITDA $6.8–7.6B (midpoint $7.2B), FCFbG $3.925–4.725B (midpoint $4.325B), with an FY2027 “opportunity” of ~$7.4–7.8B EBITDA. The “Ongoing Operations” framing (stripping Asset Closure) flatters the reported segment total by ~$75–150M/year — defensible (coal is genuinely retiring) but worth noting.

Free Cash Flow before Growth (FCFbG) — management’s headline cash metric, and the right one. FCFbG (≈ CFO − maintenance capex − cash interest − preferred dividends ± working capital, excluding growth capex) grew from $2,491M (FY23) to ~$2,888M (FY24) to ~$3.6B (FY25), guided to $3.925–4.725B (FY26). At the ~$4.3B midpoint on ~337M shares, that is ~$12.8/share of FCFbG — an ~8.6% FCFbG yield on the ~$50B equity. The critical caveat: FCFbG excludes a large, rising, majority-post-2028 growth-capex program (Vistra Zero, Permian gas, nuclear uprates). True post-growth free cash flow in build years is materially lower than the FCFbG headline implies.

Cash flow is real but lumpy — the collateral swing. Underlying GAAP CFO: –$206M (FY21), +$485M (FY22), $5,453M (FY23), $4,563M (FY24), $4,070M (FY25). The 2021–2022 near-zero/negative CFO reflects massive collateral posting during the energy-price spike (margin deposits on hedges) — a structural feature of a large hedge book that can consume cash violently when prices move against posted positions. FY25 CFO fell year-over-year partly on a +$727M working-capital use from margin deposits. Conversion of Adjusted EBITDA to FCFbG runs ~60%. (FACT — 10-K cash-flow statements.)

Balance sheet and leverage — the IG milestone. Total debt was $17,043M at YE2025 (net debt ~$15.1B; cash $785M); it jumped to ~$19,163M in Q1-2026 on Cogentrix/Lotus financing and buybacks (cash $634M, equity $5,597M). Net Debt / Adjusted EBITDA improved from 4.8x (FY21) to ~2.6x (FY25), with management targeting <3x and ~2.3x by YE2027. S&P upgraded VST to investment grade (BBB-, from BB+) in December 2025 — a genuine milestone that lowers the cost of capital and signals deleveraging discipline. Interest expense is rising ($740M FY23 → $900M FY24 → $1,179M FY25) with coverage ~5x. An 8.0% Series A perpetual preferred (~$192M/year dividend) and a Series C sit modestly atop the structure. (FACT — 10-K; Q1-2026 10-Q; S&P, Dec-2025.)

Capex — modest maintenance, rising growth. Total capex rose from $1,033M (FY21) to $2,752M (FY25). Maintenance is a modest ~$1.0–1.3B; the ramp is growth (solar/battery, Permian gas, the 433 MW of nuclear uprates tied to the Meta PPAs). Because FCFbG excludes this, the gap between FCFbG and true free cash flow widens as the growth program builds.

The share-count engine — the single biggest per-share value driver. Shares outstanding fell from 469.1M (FY21) to 338.1M (FY25) — down ~28% (–131M). Cumulatively, VST retired ~167M shares for ~$5.9B at an average cost under $36 (versus ~$148 today), with ~$1.8B of authorization remaining. The repurchases were front-loaded into the 2022–2023 trough. This is value-additive buyback — bought well below intrinsic value — not dilution-funded financial engineering. (FACT — 10-K; buyback 8-Ks.)

Quality-of-earnings flags. (1) GAAP–CFO–FCFbG divergence and violent collateral/margin swings ($406M → $1,133M of deposits posted in FY25). (2) The Asset Closure segment (retiring coal; ARO accretion $134M FY25; perennially negative EBITDA) is stripped by “Ongoing Operations.” (3) Nuclear decommissioning trust gains ($138M FY25) and asset-retirement obligations add GAAP noise and rest on critical estimates. (4) One-time items: Winter Storm Uri (2021, negative), Energy Harbor non-repeating retail gains (flatter FY25), and the Moss Landing battery fire (January-2025) — ~$400M written off via accelerated depreciation (Moss 300) plus a ~$155M impairment (Moss 100), partly offset by ~$191M insurance; only the 350 MW restarts mid-2026, and the write-offs will not recur.

ROIC/ROE — do not headline. ROE is volatile and inflated by the thin, levered equity base (see Competitive Position). GAAP operating income (FY25 $1,906M) is itself depressed by the hedge MTM swing, impairments, and Moss Landing write-offs. Use EV/Adjusted EBITDA and FCFbG/EV, not ROE or ROIC, for this business.

Verdict (financial quality): a genuine cash machine, not a capital sink — economics do improve with scale (the integrated hedge, repeatable acquisition integration, fixed-cost operating leverage) — with three honest caveats. (1) GAAP earnings are pure MTM noise; value only on Adjusted EBITDA/FCFbG. (2) FCFbG excludes a large, back-end-loaded growth-capex program, so post-growth FCF is lower than the headline. (3) FY2025 leans on non-repeating Energy Harbor/retail gains and still-elevated power prices, and the cash is commodity-/weather-/collateral-exposed and lumpy. The IG upgrade and the sub-$36 buyback are the two best financial facts in the file.

7. Capital Allocation

The per-share engine is real and well-executed. VST runs a clear waterfall: deleverage to investment grade → grow a modest dividend → opportunistic buyback → disciplined growth capex and M&A. The buyback is the crown jewel: ~167M shares retired for ~$5.9B through February-2026 at an average cost under $36 (versus ~$148 today), ~28% of the original float, with ~$1.8B of authorization remaining (the Board added $1.0B in October-2025). Front-loaded into the 2022–2023 trough — buying the stock at a fraction of its earnings power — this is the single most value-creative act in VST’s history and the textbook case for buybacks executed below intrinsic value. The dividend ($0.228/quarter, ~$0.91/year, ~15% payout of FCFbG) is small and growing; the buyback, not the dividend, is the compounding lever. (FACT — 10-K; buyback 8-Ks; Q4-2025 call.)

M&A — coherent history, pivoting to a gas roll-up at a rich moment. Dynegy (2018) added merchant gas scale out of the post-crisis bust. Ambit (2019) added retail. Energy Harbor (closed March-1-2024; ~$3.0B cash + ~15M shares + assumed debt) added 4,048 MW of PJM nuclear plus retail, made VST the #2 U.S. competitive nuclear owner, and was exceptionally well-timed ahead of the AI-power re-rating — genuinely accretive. Lotus (closed 2025; ~$1.237B + ~$800M assumed debt repaid) added ~2.6 GW gas/battery. Cogentrix (signed December-31-2025, closing 2H-2026; ~$1.9B / ~$730/kW; consideration includes ~5.0M dilutive VST shares) adds ~5.5 GW gas. Cogentrix is cheap on $/kW versus new-build (~$1,200–1,800/kW) — but it is gas-asset M&A at a cyclically high valuation for such assets amid AI-demand euphoria, financed with rising debt (total debt hit ~$19.2B in Q1-2026) and dilutive equity that partially offsets the buyback. This is the Marathon asset-growth-anomaly watch-flag — the point in the cycle where good capital allocators sometimes overpay because the theme is hot.

Growth capex. Vistra Zero (solar + battery, now impaired by OBBBA and poor ERCOT battery economics), Permian gas, and ~433 MW of nuclear uprates tied to the Comanche Peak/Meta PPAs, plus 1,200 MW of supply to an IG customer (AWS) from Q4-2027. Management cites a mid-teens levered return hurdle. A discipline flag: the 2025 annual-incentive “Total Cost” metric ran unfavorable to threshold, “primarily driven by higher spend for opportunistic growth projects” — i.e., the company is spending into the theme.

Incentive alignment — unusually good on a per-share basis, with one real gap. The DEF 14A (2026-03-18) shows: Annual Incentive weighted Adjusted EBITDA 25%, Adjusted FCFbG 25%, Total Cost 10%, Generation Operations 20%, Retail Operations 10%, Stakeholder 10% (funded 130% for 2025, with an executive limiter capping at 50% if Adjusted EBITDA misses threshold). Long-Term Incentive is 65% PSU / 35% RSU, with PSUs tied 100% to Adjusted FCFbG per share plus a relative-TSR modifier (±25% vs the S&P 500) and an absolute-TSR floor (capped at 100% if absolute TSR is negative), 200% cap. FCFbG-per-share is genuinely owner-aligned — it directly rewards the buyback and penalizes dilution. The 2023 PSUs paid the 200% maximum (FCFbG/share $24.85 vs a $17.17 target; relative TSR ~629%, ~100th percentile) — calibration is arguably soft (both the annual plan and the PSUs maxed), but the 629% TSR makes it defensible. Say-on-pay support was >97% (2025) and >98% (2024). CEO Jim Burke; independent Chairman Scott B. Helm; CFO Kris Moldovan; ~91% of CEO pay is at-risk. The real gap: there is no ROIC/ROCE gate anywhere in the incentive structure — the acquire-and-build flywheel is not explicitly return-tested in pay, and the FCFbG/share targets are undisclosed (we cannot audit their rigor). Insider ownership is thin (the ~15-person Section-16 group holds ~3.1M of 338.5M shares, <1%; Burke ~1.51M). Top holders: Vanguard 12.6%, BlackRock 8.4%, Qatar Investment Authority 5.5%, FMR 5.1%. (FACT — DEF 14A 2026-03-18.)

Insider transactions — sell-skewed and, more tellingly, no dip-buying. Across 128 Form 4s since January-2024: sales (code S) 91, grants (A) 68, tax (F) 77, option exercise (M) 73, gifts (G) 7, and open-market purchases (P) just 2 — both trivial (Director Pitesa 1,500 shares at $126.75 in March-2025; Director Crutchfield 335 shares at $89.46 in June-2024; combined ~$220K). No officer bought any stock; there was no cluster buying on the ~33% drawdown. Sales dominate (~2.17M shares / ~$353M), with CEO Burke alone selling ~789,467 shares / ~$160M — all under 10b5-1 plans, concentrated September–October-2025 at $190–215 (programmatic diversification into the run-up, not discretionary conviction selling). The read: this is the opposite of the high-conviction open-market-cluster-buy pattern we have flagged as bullish elsewhere (e.g., the UNH/ELV 2025 insider clusters). The Form 4 corpus provides no positive support to the thesis; the absence of any insider stepping in on the decline is a mildly negative tell. (FACT — Form 4 corpus, EDGAR.)

Verdict (capital allocation): competent-to-strong, entering the phase where discipline gets tested. A genuinely excellent historical allocator — the sub-$36 buyback alone made this stock for early holders — with an unusually well-aligned, per-share-based incentive plan. The asterisks: no return gate in pay, soft (maxed-out) compensation calibration, a pivot to plowing rising debt + dilutive equity into gas-asset M&A at a euphoric moment, and insiders selling (10b5-1) while none buy the dip. Strong track record; the next 18 months (Cogentrix integration, capital-cycle position) are the test.

8. Changes and Headwinds — Last Two Years

Strategic / portfolio. (1) Energy Harbor closed March-2024 — the transformational nuclear+retail acquisition that made VST the #2 U.S. competitive nuclear owner and added the East-segment nuclear now being contracted to Meta. (2) Lotus (~2.6 GW gas/battery) closed 2025. (3) Cogentrix (~5.5 GW gas, ~$1.9B) signed December-2025, closing 2H-2026 — a further gas roll-up. (4) The ~3.8 GW of 20-year nuclear PPAs with AWS (Comanche Peak, 1,200 MW) and Meta (Perry + Davis-Besse, 2,176 MW + 433 MW uprates) — the most-signed nuclear contracting of any U.S. peer, converting merchant exposure to annuity on the best part of the fleet.

Newest signal — the Helix JV (June-11-2026). VST joined Helix Digital Infrastructure, a new KKR-led company (anchor investors KKR, the Kuwait Investment Authority, NVIDIA, and Vistra; >$10B of long-duration capital committed; led by ex-AWS CEO Adam Selipsky), as both an anchor investor and the “preferred power provider” for Helix’s hyperscaler data-center buildouts — pairing NVIDIA’s DSX “AI-factory” hardware with KKR capital and Vistra generation. It is a credible validation of the data-center-power thesis and a fresh channel for Vistra to monetize its existing ~60%-utilized fleet and development pipeline at scale. But underwrite it as optionality, not earnings: the economics are undisclosed, “preferred provider” is not a contracted PPA, and Vistra’s role is a minority strategic stake. (Source: Benzinga / BusinessWire, 2026-06-11.)

Operational / one-time. The Moss Landing battery fire (January-2025) destroyed the Moss 300 facility (~$400M accelerated depreciation) and impaired Moss 100 (~$155M), partly offset by ~$191M insurance; only the 350 MW unit restarts mid-2026. This is both a financial item and a reputational/operational caution on large-scale battery deployment.

Balance sheet / ratings. S&P upgraded VST to investment grade (BBB-) in December-2025 — the culmination of the deleveraging waterfall and a structural cost-of-capital improvement.

Market / regulatory. PJM capacity auctions cleared at the cap (2026/27 and 2027/28) — a tailwind now politically contested. ERCOT completed post-Uri market-design changes (scarcity-adder redesign effective Dec-2025; PCM shelved). The OBBBA (2025) accelerated the solar/wind credit phase-out, impairing Vistra Zero economics. The ERCOT supply response (batteries doubling to ~16 GW, TEF-subsidized gas) began softening forwards — the single most important fundamental headwind.

The share-price reset. From ~$219.82 to ~$148 (~33% off the high, near the 52-week low) as the AI-power euphoria deflated. This is the dominant “change”: the market re-priced VST from a bull-case multiple back toward its base case.

Headwinds to underwrite forward. (1) ERCOT forward-price softening as supply arrives (the core bear mechanism). (2) PJM capacity clawback risk. (3) Retail normalization (~$1.6B → ~$1.4B). (4) Rising leverage from the gas roll-up into a cyclical high. (5) Natural-gas price exposure on the open ~2028+ position. (6) OBBBA/credit headwinds to the renewables build.

Verdict (changes): net neutral-to-slightly-negative for the forward thesis, despite strong backward-looking execution. The Energy Harbor deal, the nuclear PPAs, and the IG upgrade are genuine positives; but the dominant recent changes — the ERCOT supply response softening forwards, the gas-M&A pivot at a cyclical high, the OBBBA renewables headwind, and the absence of insider dip-buying — tilt the forward risk/reward more cautious than the impressive multi-year track record would suggest.

9. Risk Analysis (Risk Matrix)

Catastrophic-loss / total-loss risk: low. VST is investment grade, deeply cash-generative, and diversified across fuels, geographies, and a retail hedge; a total loss would require a simultaneous structural collapse of power prices and a refinancing crisis, which the IG balance sheet and ~$2.8B liquidity make remote. The realistic downside is a de-rating + EBITDA softening (the ~$100 bear zone), not impairment of the enterprise.

10. Valuation Discussion (Embedded Expectations)

No price target and no recommendation. Embedded-expectations and scenario analysis only.

Use the right lens. P/E and P/B are uninvestable for VST: GAAP earnings are MTM noise (see Financial Quality), and book equity is a post-bankruptcy, buyback-shrunk accounting artifact (~$5–5.6B on ~$19B debt; ~9–10x P/B means nothing). The correct lenses are EV/Adjusted EBITDA (it captures the ~$19.3B net debt that is ~27% of EV, and it is the metric management guides and segments reconcile to) and Price/FCFbG / FCFbG yield (the cash to equity and the buyback denominator; FCFbG/share is the sole PSU financial metric) — with the standing caveat that FCFbG excludes a large, back-end-loaded growth-capex program.

Current multiples (~$148; EV ~$71.7B; net debt ~$19.3B; ~337M shares; all ex-Cogentrix):

Folding in Cogentrix (~5.5 GW gas, closes 2H-2026, adds EBITDA) pulls the forward EV/EBITDA toward ~8.5–9x.

Peer comp (forward, built from guidance; trailing aggregator EV/EBITDA figures for TLN/NRG are TTM artifacts and overstated):

Placement: VST’s ~10.0x forward EV/EBITDA sits below CEG (~14x) and TLN (~11.4x), above NRG (~6.9x), and below the regulated comps. The discount to CEG is justified — VST is 82% gas+coal with ~15% nuclear versus CEG’s ~two-thirds nuclear, far less carbon-free scarcity. The premium to NRG is justified by VST’s 45U-floored nuclear, IG balance sheet, ~3.8 GW of contracted PPAs, and lower leverage. VST is a mid-price / mid-quality merchant — not a screaming relative bargain.

Own-history reconciliation — the “cheap 13.5x” is misleading. The own-10-year valuation index reads composite 77th percentile (P/E 52nd, P/B 86th, P/S 93rd). Versus its own history VST is rich, not cheap, despite the ~33% drawdown — because the drawdown compressed the multiple off a 2024–25 AI-bubble peak, not down to a distressed own-history low. P/S at the 93rd percentile and P/B at the 86th show that sales and book have not grown into the price; the EV expanded on multiple, not fundamentals. The bull must therefore argue the re-rating is durable, not that VST is cheap versus itself.

Scenario analysis (FY27–28 normalized Adjusted EBITDA × EV/EBITDA → per-share; illustrative arithmetic, NOT a target). Bridge: equity = (EBITDA × multiple) − net debt (~$19.3B → ~$21B mid-cycle), ÷ shares (~337M → ~315–325M net of buyback minus ~5M Cogentrix issuance). Cogentrix EBITDA included in all cases.

Base (~$172–181) brackets spot (~$148) modestly above; the ~33% drawdown re-priced the stock toward the base-to-bear boundary, removing the bull premium. Bull (~$245–260) approximates the prior 52-week high. Bear (~$98–105) is ~30–35% further downside. The distribution is roughly symmetric around spot, with a fat multiple-driven tail in both directions.

Embedded expectations at ~$148 / EV ~$71.7B. Reverse-engineered, at ~9.5–10x the market is capitalizing ~$7.2–7.5B of Adjusted EBITDA — essentially the FY2026 guide midpoint (ex-Cogentrix) — with little credit for FY27–28 growth, Cogentrix, or the PPA ramp. The drawdown stripped the embedded growth/re-rating premium that existed at $219.82 (where EV ~$94B implied ~$8.5–9B EBITDA at ~10–11x — the bull case priced as base).

EV decomposition (illustrative). Contracted/floored slug ~$24–34B (~35–47% of EV): ~3.8 GW nuclear under 20-year AWS/Meta PPAs (~$9–13B at ~12–14x) + 45U-floored uncontracted nuclear (~$4–7B) + ~$1.4B normalized retail (~$11–14B at ~8–10x) — annuity-like and defensible. Merchant/commodity slug ~$38–48B (~53–65%): gas+coal spreads + PJM/NE capacity + ERCOT scarcity-energy at ~7–9x — price-taking, weather-/gas-/curve-dependent, clawback-exposed. The scenario dispersion lives entirely in the merchant slug.

What the market is pricing correctly vs. incorrectly. Correctly: AI demand is real but largely priced (the de-rate removed the euphoria); the merchant half deserves a lower multiple than CEG’s nuclear book; post-drawdown, the price sits near base, not bull. Possible variant (bullish): the market gives thin value to the contracted/floored annuity slug and the IG balance sheet, treating the whole enterprise at a merchant multiple — and may under-credit the per-share buyback compounding. Possible variant (bearish): the market over-credits ERCOT energy-margin durability — energy-only, no capacity floor, post-Uri caps — if the supply flood structurally re-bases Texas forwards lower.

Sensitivities (the multiple dominates). Each 1.0x of EV/EBITDA on ~$8B ≈ ~$8B of EV ≈ ~$24–25/share; the 7.5x→11.5x range is a ~$100/share swing on flat EBITDA — dwarfing the operational levers. Natural gas: a sustained $1/MMBtu move swings the open ~half of EBITDA, but 2026–27 is hedged, so it is a 2028+ risk. ERCOT forwards: a structural re-base is the core, most-company-specific downside. PJM clawback: a multi-hundred-million-dollar EBITDA step-down risk. Whether the AI-era merchant re-rating is permanent or mean-reverts is the valuation question.

11. Variant Perception

Consensus view. The sell side is broadly constructive — VST is a favored “AI-power” / data-center-demand beneficiary with a best-in-class buyback, an IG upgrade, and ~3.8 GW of marquee nuclear PPAs; mean price targets sit well above spot (a third-party aggregate target ~$225, treated as color only, not adopted). The drawdown is widely framed as a buying opportunity in a structural-demand winner. Short interest is benign.

The strongest bull case. AI/electrification drives a durable, multi-year tightening of ERCOT and PJM; VST’s ~60%-utilized, paid-for fleet captures the demand at high incremental margins without new-build capex; the ~3.8 GW of nuclear PPAs (and more to come) convert merchant exposure to 20-year annuities at premium prices; Cogentrix and Lotus add cheap gas EBITDA into a tight firm-power market; the buyback compounds per-share value on a shrinking count; the IG balance sheet lowers the cost of capital; and the merchant multiple re-rates toward CEG-lite as the market recognizes the contracted/floored slug. → ~$245–260+.

The strongest bear case. Merchant power is a no-moat commodity at a cyclical high; the ERCOT supply response (batteries doubling, TEF gas, solar) structurally re-bases Texas forwards lower in an energy-only market with capped upside; PJM capacity is clawed back; the AI-era multiple mean-reverts to its pre-2023 high-single digits as the theme cools; FY25’s flattering retail/Energy-Harbor gains don’t repeat; and the gas-M&A pivot at a cyclical high disappoints on returns. Both EBITDA and the multiple compress at once. → ~$98–105.

The 3–5 assumptions that matter most. (1) The durability of the merchant multiple — permanent re-rating vs. mean reversion (the single biggest swing). (2) ERCOT forward power prices — does demand absorb the supply flood, or does oversupply re-base margins lower? (3) The PJM capacity windfall — held vs. clawed back. (4) Cogentrix/Lotus M&A returns — accretive vs. cyclical-top overpayment. (5) The nuclear-PPA ramp — on-time contracted annuity conversion vs. delay.

What would falsify each side. Falsifies the bull: two-plus consecutive quarters of ERCOT forward-curve softening flowing into guidance cuts, and/or a PJM capacity clawback, and/or the multiple compressing below ~8.5x on flat EBITDA. Falsifies the bear: ERCOT forwards stabilizing/firming as demand outpaces supply, additional nuclear/gas PPAs signed at disclosed premiums, and Cogentrix proving clearly accretive — converting merchant exposure to contracted annuity and validating the held multiple.

12. Fact vs. Interpretation Table

13. Open Questions

1. Pro-forma leverage and accretion after Cogentrix closes (2H-2026) — what is Net Debt/Adjusted EBITDA and the EBITDA contribution once consolidated? Current guidance excludes it, but debt already hit ~$19.2B in Q1-2026.
2. The true total growth-capex magnitude and returns — maintenance vs. growth is not cleanly disclosed; what is the all-in post-growth FCF, and do the mid-teens levered returns hold?
3. Durable retail run-rate — is ~$1.4B the right normalized number, or does easy-switch ERCOT competition erode it further?
4. ERCOT forward-curve trajectory 2028+ — does demand absorb the battery/TEF-gas/solar supply, or does oversupply structurally re-base margins? The single most important fundamental unknown.
5. PJM capacity clawback outcome — the Sept-2026 backstop auction and FERC/governor actions; how much East EBITDA is at risk?
6. 45U “gross receipts” definition — pending Treasury guidance could alter the nuclear floor’s effective level.
7. Additional nuclear/gas PPAs — will the ~3.2 GW of uncontracted nuclear (Beaver Valley, Comanche Peak) sign at disclosed premiums, validating the annuity-conversion thesis?
8. Capital-allocation discipline late-cycle — with no ROIC gate in pay and a gas-M&A pivot, does the historically excellent allocator avoid overpaying into the theme?

14. What Must Be True (Bull and Bear, with Falsification Tests)

For the BULL case (re-rating durable; ~$245–260+):

• ERCOT and PJM tighten durably as AI/electrification demand outpaces the arriving supply, holding/firming forward power prices through 2028+.
• VST signs additional nuclear/gas data-center PPAs at disclosed premiums, converting more merchant output to 20-year annuity.
• Cogentrix and Lotus prove clearly accretive; the merchant multiple holds or re-rates toward CEG-lite.
• The buyback continues compounding per-share value on a shrinking count.
• Falsification test: two or more consecutive quarters of ERCOT forward-curve softening flowing into Adjusted-EBITDA guidance cuts, OR a PJM capacity clawback, OR the multiple compressing below ~8.5x EV/EBITDA on flat EBITDA — any of these breaks the bull.

For the BEAR case (de-rate + EBITDA softening; ~$98–105):

• The ERCOT supply flood (batteries doubling, TEF gas, solar) structurally re-bases Texas forwards lower in an energy-only, capped market, compressing the ~46% Texas slug.
• PJM capacity is clawed back; the AI-era multiple mean-reverts to pre-2023 high-single digits.
• FY25’s flattering retail/Energy-Harbor gains don’t repeat; the gas-M&A pivot disappoints on returns.
• Falsification test: ERCOT forwards stabilizing or firming as demand outpaces supply, AND additional nuclear/gas PPAs signed at premiums, AND Cogentrix proving accretive — any sustained combination of these breaks the bear and validates the held multiple.

The Source Appendix follows below (Appendix B).

APPENDIX A — Standard Diligence Questionnaire

Vistra Corp. (NYSE: VST) — as of 2026-06-13

Supplemental to the memo. Fact/Interpretation/Assumption labels applied where material.

General

What thoughtful questions have other investors asked about this company?

• Is the AI/data-center power-demand re-rating permanent or a cyclical/thematic bubble that mean-reverts? (The single most-debated question.)
• How much of VST’s earnings power is genuinely contracted/floored (nuclear PPAs + 45U + retail) versus merchant/commodity? (Roughly 35–47% vs 53–65% of EV.)
• Is ERCOT’s energy-only market — with no capacity floor and a fast-arriving battery/gas supply response — a structural problem for VST’s largest (~46%) geography?
• Is VST overpaying for gas assets (Cogentrix, Lotus) late in the capital cycle?
• Why is GAAP EPS so volatile, and which metric should one actually use? (Adj EBITDA / FCFbG, not GAAP.)
• How durable is the retail book in a market designed for easy switching?

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Closer to a cyclical high (Interpretation). Power prices recovered sharply off the 2021 trough; PJM capacity cleared at the auction cap; ERCOT forwards are now softening as supply arrives. The own-history valuation sits at the 77th percentile (P/S 93rd). 2026–27 is heavily hedged, so cyclical risk is concentrated in 2028+.

Driven by the external environment or internal actions? Both. External: post-Uri power-price recovery, the AI-demand window, PJM capacity prices. Internal: the Energy Harbor acquisition (added ~$2B+ EBITDA), operating execution, and the ~28% share-count reduction that drove per-share value.

How stable are revenues? Reported revenue is volatile, gross-settled, and pass-through-laden — a poor stability gauge. Adjusted EBITDA has risen every year FY21→FY25, smoothed by the integrated gen+retail hedge (Q1-2026 proved the hedge: mild weather hurt retail, generation offset it). But the underlying merchant margins are commodity-, weather-, and forward-curve-dependent.

Outlook for products/services? Electricity demand growth is genuinely robust (~5–6%/year ERCOT per management, vs ~1–2% historically) — a real tailwind. The risk is on the supply/price side (oversupply compressing margins), not demand.

How big will this market be — growing, shrinking, domestic or international? Domestic (US: ERCOT, PJM, ISO-NE, NYISO, MISO, CAISO). The end market (US power demand) is growing for the first time in ~two decades, driven by data centers, electrification, and reshoring. But it is a commodity market where price, not volume, determines profitability.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More, on the supply side — TEF-subsidized new gas, merchant batteries doubling to ~16 GW, and surging solar are all entering ERCOT, the Marathon capital-cycle signature. Merchant power has zero entry barriers.

How profitable is the business (ROIC, ROE)? Headline ROE ~43% is not a reliable signal — it is inflated by a thin, post-bankruptcy, buyback-shrunk equity base and cyclically peak prices. Through-cycle, returns are far more pedestrian. Use EV/Adj-EBITDA (~10x fwd) and FCFbG yield (~8.6%), not ROE/ROIC.

How profitable is the industry — competitors, barriers to entry? Structurally a bad industry (Greenwald): a commodity with no barriers and mean-reverting returns, currently enjoying a demand windfall. Competitors: CEG, Talen, NRG, Calpine (now CEG), regulated utilities. No durable industry profit pool absent the current demand cycle.

Can the business be easily understood? Moderately. The asset base and retail book are simple; the earnings are obscured by commodity-hedge MTM (GAAP is noise) and require translating to Adj EBITDA/FCFbG.

Can it be undermined by foreign low-cost labor? No — power is generated and consumed locally; this is not a labor-arbitrage-exposed business.

Do brands matter? Modestly. TXU is a ~20-year brand with ~2.6M Texas customers, providing weak demand captivity (habit/inertia) — but in a market engineered for easy switching, it is not a strong moat.

What is the nature of competition? On generation: price competition on a commodity (marginal-cost dispatch). On retail: price + brand + service in a low-switching-cost market.

Customers’ switching costs? Low (generation: none — the grid buys the cheapest MWh; retail: deliberately low under ERCOT Power-to-Choose).

Moat verdict: No durable franchise moat. The only financializable edges are the narrow nuclear scarcity/45U floor (~15% of capacity) and a demonstrated capital-allocation/acquire-integrate capability (not a barrier). “Best operator in a bad industry.”

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The ~3.8 GW of 20-year nuclear PPAs (AWS/Meta) are contracted annuity value not capitalized as such; the irreplaceable nuclear licenses; the TXU brand. Conversely, the equity base understates economic value due to the bankruptcy-fresh-start accounting and ~$6.9B treasury stock.

Off-balance-sheet liabilities? Operating leases, fuel/transport contracts, and — material to this business — large commodity-hedge collateral obligations that can consume cash in price spikes (CFO turned negative in 2021–22). Nuclear decommissioning and asset-retirement obligations are on-balance-sheet but rest on critical estimates.

How conservative is the accounting? Mixed. Management’s non-GAAP presentation (Adj EBITDA, FCFbG, “Ongoing Operations”) is reasonable and the right lens, but “Ongoing Operations” flatters the total by ~$75–150M/year by stripping Asset Closure. GAAP is dominated by MTM noise. No evidence of aggressive revenue recognition.

How CapEx-hungry is the business? Maintenance capex is modest (~$1.0–1.3B), but the growth program (solar/battery, Permian gas, nuclear uprates) is large and back-end-loaded (majority post-2028). FCFbG excludes growth capex, so post-growth free cash flow is materially lower than the headline.

Capital Allocation & Management

How much FCF does the business generate, and how is it used? FCFbG ~$3.6B (FY25), guided $3.9–4.7B (FY26). Used per a waterfall: deleverage to IG → modest growing dividend (~15% payout) → opportunistic buyback (~$5.9B / 167M shares at <$36) → disciplined growth capex and M&A.

Significant acquisitions recently? Energy Harbor (nuclear, Mar-2024), Lotus (gas, 2025), Cogentrix (~5.5 GW gas, ~$1.9B, closing 2H-2026). The early deals were well-timed in distress; the recent gas roll-up is cheap on $/kW but at a cyclically high moment (Marathon asset-growth flag).

Buying back shares? Yes, aggressively and value-accretively — ~28% of float retired at avg <$36. The single biggest per-share value driver.

Issuing large amounts of new shares to insiders? No large dilutive insider issuance; the ~5M Cogentrix-deal shares modestly offset the buyback. SBC is modest.

Compensation policy? Well-aligned on a per-share basis: annual incentive weights Adj EBITDA 25% / Adj FCFbG 25% / operations / stakeholder; LTI is 65% PSU tied 100% to Adjusted FCFbG per share + relative-TSR modifier + absolute-TSR floor. Say-on-pay >97%. Gap: no ROIC gate; targets undisclosed; 2023 PSUs paid the 200% max (soft calibration, though a 629% TSR makes it defensible).

Motivations of management? Per-share value creation, evidenced by the buyback and FCFbG/share PSU. But insider ownership is thin (<1%), and the Form 4 record shows sales (all 10b5-1, ~$353M incl. CEO ~$160M) and no dip-buying on the ~33% drawdown — no conviction-buy signal.

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer? No — a standard US C-corp common share (NYSE: VST); 1099, not K-1.

Dividend policy? Modest and growing: ~$0.91/year (~$0.228/quarter), ~15% payout of FCFbG, ~0.62% yield. The buyback, not the dividend, is the return lever. An 8.0% Series A perpetual preferred sits atop common.

How profitable is the business? On a cash basis, genuinely profitable: ~$5.84B Adj EBITDA (FY25), ~$3.6B FCFbG, ~60% conversion. On a GAAP basis, volatile and uninformative (MTM noise).

Is net income diverging from cash from operations? Yes, structurally — GAAP net income is MTM-driven and diverges sharply from CFO/FCFbG. This is the defining QoE feature: value on cash, not GAAP earnings.

Risks & Downside

What factors would cause the stock to decline? ERCOT forward-price re-base lower (supply flood); multiple mean-reversion (the biggest swing); PJM capacity clawback; natural-gas collapse (2028+); cyclical-high M&A disappointing; retail normalization; a data-center-demand shortfall; an operational event (another Moss-Landing-type fire).

Risk of a catastrophic loss? Low. Investment grade (BBB-), deeply cash-generative, diversified across fuels/geographies/retail, ~$2.8B liquidity. The realistic downside is a de-rate + EBITDA softening (the ~$100 bear zone), not enterprise impairment.

Chance of a total loss? Very low — would require simultaneous structural power-price collapse and a refinancing crisis, remote given the IG balance sheet and cash generation.

Recent News & Events

Has the business environment changed recently? Yes: (1) the AI-power euphoria deflated — the stock fell ~33% from $219.82 to ~$148; (2) the ERCOT supply response (batteries doubling to ~16 GW, TEF gas) began softening forwards; (3) S&P upgraded VST to investment grade (Dec-2025); (4) OBBBA (2025) accelerated solar/wind credit phase-outs, impairing Vistra Zero.

Significant acquisitions? Cogentrix (~5.5 GW gas, signed Dec-2025, closing 2H-2026); Lotus (2025); Energy Harbor (Mar-2024). ~3.8 GW of nuclear PPAs signed with AWS and Meta.

Change in accounting policies? No material change; management began emphasizing a simplified two-bucket “Ongoing Operations” (Generation + Retail) presentation.

Recent changes — new markets, facilities, management? The Energy Harbor nuclear added PJM/Ohio exposure; the Moss Landing battery fire (Jan-2025) removed capacity; CEO Jim Burke and CFO Kris Moldovan lead; the company continues its gas roll-up and nuclear-PPA contracting strategy.

APPENDIX B — Source Appendix

Vistra Corp. (NYSE: VST) — as of 2026-06-13

Primary sources first. Quantitative figures are reconciled to SEC filings; third-party market data is treated as convenience/cross-check only.

A. SEC Filings (primary — EDGAR, CIK 0001692819)

B. Earnings Call & Event Transcripts (company IR / public transcripts)

C. Quantitative Data Helpers (convenience; reconciled to filings)

D. Industry, Regulatory & Market Sources (public)

E. Peer Cross-Read (public filings)

Note on data sources: third-party market-data aggregations are treated as convenience/cross-check and reconciled to SEC filings for all material numbers. Analyst price targets (aggregate ~$225) are noted as third-party color only and are not adopted as a target.