Power & utilities · Coverage initiation · October 3, 2026

Vistra: The Street Has the Earnings. It Doesn’t Have the Rate.

Author Brandon Leon Posted Saturday, October 3, 2026, with the stock at $140.02 (Oct 2 close) Coverage Initiation · HOLD · 12-to-18-month price target $140 · four lenses in §5 · proof gate at the Q3 report, November 6

TL;DR: We are initiating coverage of Vistra at HOLD with a $140 price target, roughly where the stock closed Friday. Vistra is the largest competitive power generator in the country wrapped around the largest retail electricity book, and the business is in better shape than a 36% drawdown suggests: EBITDA grew 31% in the second quarter, 2026 guidance held, and 94% of 2027 output is already sold forward. We agree with the Street on 2026 and 2027 earnings to within 1%. Where we part ways is what those earnings are worth. With the 10-year Treasury at 5.26%, our discount rate is 8.7% and our base-case DCF is $144; the Street's $218 average target requires about 6.5%, and even on the Street's own EBITDA our model reaches only $161. The open question is 2028, the first year Texas forward curves really bite, and the consensus number for it sits on our optimistic path. We would buy below about $122. No position.

What’s in the box This is the web edition of a full initiation written in the CFA Institute Research Challenge format: an 18-page report with 37 exhibits and five appendices, a formula-driven three-statement model with a ten-year DCF across three scenarios, and a one-page tear sheet. All three are free to download at the end of this note. The model is priced at the September 30 close of $138.35; the stock has since moved to $140.02, which changes the arithmetic by about a dollar and the conclusion not at all.

1. What Vistra actually sells: hedged power

Vistra owns about 44 GW of generation — 62% natural gas, 20% coal and lignite, 15% nuclear — across Texas (ERCOT), the Mid-Atlantic (PJM), New York and New England, and about 49 GW once the $4.0 billion Cogentrix acquisition closes later this year.⁴ It also sells electricity to about 5 million retail customers under TXU Energy and a handful of other brands. That combination is the business. When wholesale prices spike, retail margins get squeezed and the generation fleet earns more; when prices collapse, the reverse. The cleanest demonstration was Winter Storm Uri in 2021: retail earned $1.3 billion while Texas generation lost $236 million, and the company survived a week that bankrupted competitors.³

On top of the natural hedge, management sells generation forward two to three years. As of August 3, about 100% of expected 2026 output was hedged, 94% of 2027 and 72% of 2028.¹ That is the single most important fact for anyone reading the stock chart: the forward-curve decline that has driven the shares lower cannot touch reported earnings for at least five quarters.

The newer layer is contracted nuclear. Comanche Peak signed 1,200 MW for 20 years with Amazon Web Services, energizing in late 2027; three PJM plants signed 2,609 MW for 20 years with Meta, including 433 MW of uprates.⁵ Together they put more than half of the operating nuclear fleet under investment-grade contracts, and the federal nuclear production tax credit puts a floor of roughly $44/MWh under most of the rest through 2032.

2. The numbers that are real

Ongoing Operations Adjusted EBITDA tripled from $1.94 billion in 2021 to $5.91 billion in 2025, through the Energy Harbor nuclear acquisition, PJM capacity prices that went from $29 to $270 per MW-day, and hedges rolling onto the higher 2023–25 curves. The second quarter of 2026 printed $1.77 billion, up 31%, and guidance of $6.8–7.6 billion was reaffirmed, with management saying free cash flow will land at or above the midpoint.¹ Since November 2021 the company has bought back about $6.5 billion of stock and cut its share count by roughly 30%.

Chart 1 — Four years, three times the EBITDA

The earnings base kept compounding while the stock fell 36%.

FY21-FY25 as originally reported; FY26 is the guidance midpoint and range; FY27E is our base case including Cogentrix (about $560M), in line with the $8.05B consensus.¹

Our forecast has $7.20 billion of EBITDA in 2026 and $8.04 billion in 2027, both within 1% of consensus. Free cash flow before growth is about $4.3 billion this year, or $12.70 a share — a 9.1% yield on Friday’s close. If this were the whole story, the stock would be cheap.

3. The number that moved: the Texas curve, and 2028

Between the third-quarter 2025 and second-quarter 2026 reporting dates, ERCOT North Hub forwards fell 21% for 2027 and 15% for 2028, while PJM West rose 22%.⁶ Record Texas demand — an unofficial 91.1 GW peak on July 22 — did not produce record prices, because batteries now shave the evening peaks that used to set generator margins. Management’s example: real-time prices settled near $57/MWh at record load, where it believes they could have reached $400–500 without storage.² Then on August 3 Texas paused data-center interconnections pending an audit of a 474 GW queue, with a report due December 10.⁸

Chart 2 — The curve that sets the hedges

Texas forwards fell. PJM forwards rose. Vistra owns both.

Around-the-clock forwards from Vistra's Q3'25 (as of 9/30/25) and Q2'26 (as of 6/30/26) results presentations.⁶

Because Vistra hedges forward, the curve reaches earnings with a lag, and 2028 is the first year it lands in size: only 72% hedged, with much of that hedged after the curve rolled over. PJM capacity prices at the $325–333/MW-day cap offset part of it, but only through May 2029, and on September 30 FERC unexpectedly delayed PJM’s backstop procurement over cost-allocation concerns.⁷ Our base case has 2028 EBITDA at $8.19 billion, 7% below the $8.79 billion consensus. The consensus figure sits on our optimistic path, which assumes a second Texas hyperscaler contract and a recovery in ERCOT prices.

Chart 3 — Agree on 2026-27, disagree on 2028

Consensus 2028 is our optimistic case.

Our three scenarios from the model; consensus from MarketScreener as of September 30.⁹

Where we’d be wrong on 2028

Management calls the Texas curve “recency bias” after two mild summers and notes that battery returns have fallen to about one-fifth of initial expectations, which should slow new storage. If a hot summer or a stalled battery pipeline lifts ERCOT forwards back toward the $55–60/MWh of late 2025 before the 2028 book is filled, our 2028 number is too low and consensus is right. A second 1+ GW Texas data-center contract after the December audit would get there faster. Either is the path to our $173 optimistic case, and neither requires anything heroic.

4. The price, and what the Street is paying for

The stock closed $140.02 on Friday, 36% below its 52-week high of $217.10, a market value of about $47 billion. Nineteen of twenty analysts rate it a Buy; the average target is $218, with a range of $106 to $305.⁹ The CEO bought about $1.2 million of stock near $135 in late August, after years of selling at $180–210.¹¹ On the surface this is a classic setup: a quality business, a big drawdown, insiders buying and a Street that sees 55% upside.

The question we kept returning to is what changed between the 2025 peak and now. Earnings went up. What went up faster is the 10-year Treasury, to 5.26%, with the 30-year at its highest since 2002.¹⁰ Power stocks were priced in 2025 as scarcity assets with bond-like contracts attached. Bond-like assets reprice when bonds do.

5. Four lenses, one number

We value the business four ways and publish all four, in the order of the weight we give them.

Lens 1 — Multiples, against the IPP group

Pro forma for Cogentrix, Vistra trades at 9.2× our 2027 EBITDA. That is below Constellation (12.1×), whose nuclear-heavy fleet earns its premium, in line with Talen (8.9×) and above NRG (7.2×), which is down 42% in a year. Applying the 9.4× average of the three listed IPPs gives $146. A free-cash-flow yield method — 2027 free cash flow after growth capex and preferred dividends, capitalized at 7% — gives $139.¹²

Chart 4 — Priced between the poles

Vistra sits between Constellation's nuclear premium and NRG's discount.

EV and consensus EBITDA as of September 30; VST pro forma for Cogentrix consideration and our 2027E EBITDA.⁹

Lens 2 — Discounted cash flow

A ten-year unlevered DCF from Q4 2026 through 2035, mid-year convention, terminal value blending a 9.0× exit multiple with 2.5% perpetual growth. The inputs worth knowing: a 1.32 beta (peer betas unlevered, averaged and relevered at Vistra’s pro forma structure), Damodaran’s 4.14% implied equity risk premium, a 6.2% pre-tax cost of debt, and a 8.7% WACC. Cash taxes rise from 4% this year to the 21% statutory rate by 2029 as loss carryforwards run out, which is the largest single drag on free-cash-flow growth. The equity bridge deducts $19.9 billion of debt, $2.5 billion of preferred and the $3.8 billion of Cogentrix cash and assumed debt.¹²

DCF case	EBITDA 2027E → 2028E	Probability	Per share
Pessimistic	$7.67B → $7.23B; Texas curve right, PJM capacity normalizes after 2029	25%	$93
Base	$8.04B → $8.19B; no Texas deal beyond AWS, PJM flat after the cap	50%	$144
Optimistic	$8.29B → $8.87B; second Texas hyperscaler deal, ERCOT recovers	25%	$173
Probability-weighted	—	100%	$139

Lens 3 — Reverse DCF: what $218 already assumes

Hold our base-case cash flows fixed and solve for the Street. The $218 average target requires either a WACC of about 6.5% at our 9.0× exit, or a 15× exit multiple in 2035 at our 8.7%. At Vistra’s capital structure, 6.5% implies a cost of equity of about 7% — less than two points over a 5.3% Treasury for a stock with a beta above 1.3. A 15× terminal multiple would value a 2035 fleet of gas and aging nuclear plants like a regulated utility.

Chart 5 — Same cash flows, different rate

The Street's target is a discount-rate call, not an earnings call.

Base-case DCF from the model at a 9.0x exit multiple across discount rates; Street mean from stockanalysis.com.¹²

Then the test that matters most: we re-ran the DCF on consensus EBITDA instead of ours — $7.23 billion, $8.05 billion and $8.79 billion for 2026–28, held above our path after that. It gives $161. That is $17 above our base case and still $57 below the Street’s target. We largely agree with the Street on the earnings. We disagree on what they are worth at a 5.3% risk-free rate.

Steelmanning the Street — because someone has to

If long rates fall back toward 4.5%, our WACC drops toward 8% and the base case is worth about $163 before anything changes in Texas. Add a second hyperscaler contract and the optimistic case at that rate is worth about $195. That is a coherent view, and it is how most of the twenty analysts get to their targets. It is also two bets — on rates and on a contract — stacked on top of a business we already like. We would rather pay for one of them than both.

Lens 4 — The weighting, and our number

Method	Value	Weight
DCF, probability-weighted	$139	50%
EV / 2027E EBITDA at the 9.4× IPP average	$146	25%
Free-cash-flow yield at 7.0% (2027E)	$139	25%
P/E at the 14.5× peer median (cross-check)	$145	—
Price target	$140	rounded from $141

The weighted fair value is $141, which we round to a $140 price target — essentially where the stock closed on Friday. We initiate at HOLD. Every method lands within about 6% of the price, which is the point: the market has done the arithmetic. What it has not yet done is price the skew. Our pessimistic case is $47 a share below the stock and our optimistic case $33 above it, and the pessimistic case requires nothing more than the forward curve being right.

6. The framework

We initiate at HOLD, about 36% below the Street’s average target. Same posture as Bloom: consensus on the business, contrarian on the price. The difference here is that the business case is better and the price is fair rather than stretched, so the levels below are about when to act, not when to get out.

Chart 6 — Thirty-three months, with the levels drawn on

From $41 to a $209 close and back to $140. The levels are where discipline lives.

Monthly closes from Yahoo Finance; the last point is the October 2 close. Levels are ours, from the model.⁹

Level	What it means
Below ~$122	Our target would sit 15% above the stock: upgrade to BUY, provided the thesis is intact (the 2027 guide at or above $8.1 billion; no deterioration in 2028 hedge prices).
Near $93	The market is pricing our pessimistic case; 2026–2030 free cash flow before growth would cover roughly 70% of the market value.
Above $173	The price exceeds our optimistic case. We would move to SELL unless a second Texas contract or lower long rates had lifted our base case first.
Thesis breaks (HOLD → AVOID)	2028 ERCOT hedges placed below $45/MWh · a PJM reform that lowers capacity prices after May 2029 · a leveraging acquisition before 2028 · Moody’s moving away from investment grade.
The one number	The 2027 outlook with third-quarter results on November 6. At or above $8.1 billion including Cogentrix and Meta confirms our base case; above $8.3 billion puts the optimistic path in play.

Last, the caveat that keeps this honest. This site’s readers know our book leans toward the AI-infrastructure factor, and Vistra trades with it: it fell 5% in the September 14 AI-power selloff on no company news.⁹ Owning it adds a name, not diversification. That is one more reason the rating is HOLD with a level, not a chase.

Simply: A well-run power company, hedged through 2027, with real contracts and a buyback that works. The Street has the earnings right. At today’s interest rates the stock is worth about what it costs, so we rate it HOLD at $140 and would buy below $122. The next test is the 2027 outlook on November 6.

Full report (PDF, 18 pages) → Model (.xlsx) One-pager

Get the next note by email · All writing

Sources & footnotes

Results and guidance: Vistra Q2 2026 release, August 7, 2026 — Ongoing Operations Adj. EBITDA $1,767M (+30-31%), H1 $3,261M; 2026 guidance $6.8-7.6B EBITDA and $3.925-4.725B FCFbG reaffirmed; 2027 “midpoint opportunity” $7.4-7.8B excluding Cogentrix and Meta; hedged ~100% / 94% / 72% for 2026-28 as of August 3; ~$6.5B of buybacks since November 2021 (Vistra IR). ↩
Q2 2026 earnings call, August 7, 2026: 2027 trending to the low end on ERCOT curves “meaningfully lower than October 31st, 2025”; July 22 ERCOT pricing and battery commentary; FCFbG “at or above midpoint” (transcript). ↩
FY2025 results (February 26, 2026) and 10-K: Adj. EBITDA $5,912M; fleet of 43,641 MW at year-end 2025 by fuel; segment history including 2021 Retail $1,312M and Texas ($236)M (Vistra IR). ↩
Cogentrix: ~$4.0B net ($2.3B cash, ~5M shares, $1.5B assumed debt, less ~$0.7B tax benefits), 5.5 GW across PJM, ISO-NE and ERCOT, ~7.25x 2027E EBITDA; FERC approval August 2026; close expected late 2026 (release). ↩
Contracted nuclear: Meta, 2,609 MW over 20 years from Perry, Davis-Besse and Beaver Valley, including 433 MW of uprates (release); AWS, up to 1,200 MW over 20 years at Comanche Peak, named on the Q4'25 call. Section 45U nuclear PTC preserved through 2032 in the July 2025 tax law. ↩
Forward curves: Vistra Q3 2025 and Q2 2026 results presentations, around-the-clock forwards as of September 30, 2025 and June 30, 2026 — ERCOT North 2027 $58.83 to $46.26, 2028 $58.88 to $50.11; PJM West 2027 $54.72 to $66.51 (Q2 deck). ↩
PJM capacity: RTO clearing prices $329.17 (2026/27), $333.44 (2027/28), $325.00 (2028/29), the last ~6,831 MW short of the reliability requirement (PJM); FERC's delay of the Reliability Backstop Procurement, September 30, 2026 (Investing.com). ↩
Texas: Governor Abbott's August 3 directive pausing large-load energization pending an audit of ERCOT's ~474 GW queue, report due December 10 (Utility Dive); ERCOT unofficial peak of 91.1 GW on July 22, 2026. ↩
Market data: VST closes from stockanalysis.com and Yahoo Finance (Oct 2 close $140.02; 52-week range $132.66-$217.10); 20 analysts, mean target $217.58, range $106-$305 (stockanalysis.com); consensus EBITDA and EPS from MarketScreener; peer EV and multiples as of September 30; September 14 sector selloff per TIKR. ↩
Rates: 10-year Treasury par yield 5.26% and 30-year 5.59% on September 29, 2026 (US Treasury); Damodaran implied equity risk premium 4.14% (September 1, 2026). ↩
Insider purchases: CEO Jim Burke bought about 8,665 shares at ~$135-136 in late August 2026, about $1.17M, after prior sales (Motley Fool, Form 4). ↩
Valuation: all four lenses from the desk’s VST model (downloadable above). WACC 8.70% = cost of equity 10.73% (rf 5.26%, beta 1.32, ERP 4.14%) at 64% weight, after-tax debt 4.90% at 33%, hybrids 7.10% at 3%. DCF: Q4'26-2035 FCFF, mid-year, TV = 50% x 9.0x 2035E EBITDA + 50% Gordon at 2.5% (TV = 56% of EV). Equity bridge: 6/30/26 funded debt $19,895M, preferred $2,476M, Vistra Vision obligation $613M, Cogentrix cash and assumed debt $3,800M, less cash $435M and $700M of acquired tax benefits; 340.6M shares. Reverse DCF: Street mean at our WACC needs a 15.1x 2035 exit (100% exit TV); at 9.0x, ~6.5% WACC. Consensus-EBITDA DCF: $161. Pricing in the model is the September 30 close ($138.35). ↩

Methodology: tests are graded against their written wording. Quotes are verbatim where primary documents or official transcripts are available; other factual claims are sourced in the notes. Fills and positions are disclosed only when real.

Nothing on this page is investment advice. We work in scenarios and trip-wires, not predictions. See disclaimer.

Disclosure: I/we have no position in Vistra Corp. (VST) and no plans to initiate one within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. This commentary is for informational and educational purposes only and does not constitute investment, tax, or legal advice. Past performance is not indicative of future results. See disclaimer.