How Much Power Will Be Enough?
A regional shortage is forming in the U.S. grid. The equipment makers helped reveal it — but the stronger near-term financial opportunity may be in the companies already monetizing scarce, dispatchable generation.
One scarcity thesis. Two financial expressions.
VST and TLN are correlated exposures, not diversification from one another.
More protected
Heavier hedging, broader geographic exposure, long-duration contracts and aggressive capital return reduce how much of the next twelve months depends on tomorrow’s power price.
More convex
Greater PJM concentration and a more open 2028 hedge book create more sensitivity to continued scarcity, new data-center contracting and forward power economics.
The equipment makers revealed the shortage. They did not necessarily reveal the best stock.
The U.S. power story has been told mostly through equipment.
Gas turbines are booked years in advance. Transformer lead times remain measured in years. Electrical-equipment manufacturers are reporting large order books. Data centers, industrial expansion, electrification and replacement of aging grid infrastructure are all competing for the same limited physical system.
That evidence matters. It tells us the scarcity is real.
But scarcity alone does not identify the best stock.
MAi Edge Capital's meta-analysis moved one step further: where is the scarcity already producing cash, and where has the market not fully paid for it?
That distinction leads away from the most obvious hardware beneficiaries and toward two owners of existing dispatchable generation: Vistra Corp. (VST) and Talen Energy (TLN).
The central conclusion is not that the equipment companies are wrong. GE Vernova, Eaton, Quanta, nVent and others have real backlogs and real demand. The issue is price. Much of the equipment scarcity is already reflected in elevated valuations.
Vistra and Talen sit at a different point in the chain. They already own power plants. They are already selling electricity and capacity into constrained regions. Their cash flows are partially hedged or contracted. And their valuations still reflect substantial concern that regulators and politicians will prevent the owners of scarce generation from capturing the full economics of that scarcity.
That discount is the opportunity.
The shortage survived the price cap
The clearest evidence comes from PJM, the regional grid operator serving more than 65 million people across parts of the Mid-Atlantic and Midwest.
PJM's 2028/2029 capacity auction cleared at the FERC-approved cap of $325 per MW-day. Even at that capped price, the system procured 6,831 MW less capacity than its reliability requirement.
That combination matters.
The price signal was suppressed by policy — and the physical shortage remained.
PJM reported that the auction secured 138,318 MW of unforced capacity and demand response, with the total available supply still falling short of the amount required to satisfy its one-event-in-ten-years reliability standard.
The market therefore is not asking whether scarcity exists. It is asking how much of the scarcity value generators will ultimately be allowed to keep.
That is a different question.
Political intervention is already in the price
The current PJM capacity regime already includes a price collar. The $325-per-MW-day cap is not a hypothetical future risk. FERC accepted PJM's tariff revisions extending the collar into the relevant auctions in April 2026.
That changes the investment framing.
The thesis is not that regulation will disappear. It is that Vistra and Talen are being valued under an intervention regime that already exists, while a meaningful portion of their near-term cash flow is protected by hedges, long-term contracts and existing generation assets.
PJM is also pursuing a Reliability Backstop Procurement, or RBP, to address the 6,831 MW shortfall. Its July proposal contemplated contracts as long as 15 years, with maximum willingness to pay based on a MW-weighted average of roughly $555 per MW-day for new supply.
That creates an unusual contrast: existing capacity clears under a $325 cap while policymakers are considering paying materially more to induce new capacity.
The RBP is two-sided for investors.
Over time, new generation can reduce scarcity rents. But new gas generation cannot be created on command. Turbine slots, transformers, interconnection, permitting and construction all take years. The market and valuation impact of an RBP approval could arrive quickly, but its ability to add large amounts of physical dispatchable supply is much slower.
That gives existing generators a window in which the shortage can continue to produce cash before new capacity materially changes the balance.
H.R. 9340 does not break the thesis
The House-passed Ratepayer Protection Act, H.R. 9340, is another example of policy responding to the power-demand surge.
The legislation is aimed primarily at cost allocation. It would establish a federal standard requiring regulators to consider whether very large new loads should bear the full incremental cost of generation, transmission and distribution upgrades needed to serve them, along with financial assurances against project termination.
That is not a direct cap on PJM capacity prices.
For the scarcity thesis, it may actually improve the quality of the demand signal. Requiring large projects to fund the infrastructure they trigger should remove some speculative or duplicate load requests while leaving economically real projects in the queue.
For VST and TLN, the more important political risk remains what comes after the current collar: whether price suppression extends beyond the 2029/2030 delivery year, whether backstop procurement adds supply faster than expected, or whether future market reforms materially reduce merchant-generation economics.
The more protected expression
Vistra is the primary MAi expression of the thesis because more of its near-term economics are already protected.
For the second quarter of 2026, Vistra reported $1.767 billion of Ongoing Operations Adjusted EBITDA. The company reaffirmed full-year 2026 guidance of $6.8 billion to $7.6 billion of adjusted EBITDA and $3.925 billion to $4.725 billion of adjusted free cash flow before growth.
Just as important, its expected generation volumes were approximately 100% hedged for 2026, 94% for 2027 and 72% for 2028 as of early August.
Those hedges matter because they reduce the amount of the near-term thesis that depends on tomorrow's power price.
Vistra has also been steadily shrinking its share count. Since 2021, the company has retired roughly 30% of its shares through repurchases and continues to return capital aggressively.
Long-duration commercial contracting adds another layer. Vistra has announced a 20-year power purchase agreement with Amazon Web Services tied to up to 1,200 MW at Comanche Peak, along with long-term Meta agreements covering roughly 2,600 MW across four PJM nuclear facilities.
There is counterevidence, and it matters.
Management has indicated that the 2027 opportunity is trending toward the lower end of its prior range as ERCOT forward prices softened. And Lone Pine exited a large VST position during the second quarter.
Against that, CEO Jim Burke made open-market purchases in August and September — the most notable insider-buying signal identified in the broader power-sector screen.
That is exactly the kind of contradiction the MAi process is designed to surface. The conclusion is not that Vistra is riskless. It is that the market may be discounting the political and commodity risks more heavily than the company's hedged and contracted cash-flow profile warrants.
MAi 12-month distribution — VST
Reference price: $138.46 at the Sept. 25, 2026 close.
| Branch | Probability | Analytical range | What produces it |
|---|---|---|---|
| Bear | 30% | $110–120 | More political extraction beyond the existing collar, RBP expectations pressure forward value, ERCOT pricing remains weak, 2027 lands near the low end of management's opportunity range. |
| Base | 45% | $165–180 | Current collar remains the principal constraint, RBP adds supply slowly, hedged cash flows convert, PPAs and buybacks continue. |
| Bull | 25% | $210–225 | Additional large-load contracting, stronger realized power economics, better-than-expected 2027 earnings and continued capital return. |
The more convex expression
Talen represents the same basic scarcity factor, but with more open exposure to what happens next.
The company raised its 2026 outlook after the second quarter. Revised guidance called for $2.025 billion to $2.225 billion of adjusted EBITDA and $1.20 billion to $1.35 billion of adjusted free cash flow.
Talen's hedge profile is materially more open than Vistra's: approximately 85% of expected generation volumes were hedged for 2026, 70% for 2027 and only 30% for 2028 at June 30.
That is the central difference between the two stocks.
Vistra has more protection. Talen has more optionality.
Talen's own forward framework showed adjusted free cash flow per share around $26.60 for 2026, approximately $34 for 2027 and approximately $40 for 2028 before giving full credit to the company's additional upside levers.
Its 2028/2029 PJM capacity position is also significant. Talen cleared 10,180 MW at the $325-per-MW-day cap, equivalent to roughly $1.208 billion of capacity revenue for the delivery year before considering other elements of the generation portfolio.
The company also has substantial remaining repurchase capacity, creating a direct mechanism for converting free cash flow into per-share value.
But because more of the 2028 book remains open, TLN carries more sensitivity to both sides of the thesis. If scarcity persists, forward power economics and new data-center contracting can produce more upside than at Vistra. If policy, new supply or weaker power prices arrive faster than expected, more of that exposure is available to move against shareholders.
MAi 12-month distribution — TLN
Reference price: $304.42 at the Sept. 25, 2026 close.
| Branch | Probability | Analytical range | What produces it |
|---|---|---|---|
| Bear | 30% | $240–260 | Longer-lived political suppression, faster RBP supply response, weaker forward power economics and a markdown of TLN's more open 2028 position. |
| Base | 45% | $360–400 | Current policy framework remains intact, cash-flow conversion follows the company's forward framework, repurchases continue and new supply arrives slowly. |
| Bull | 25% | $450–500 | Stronger spark spreads, material powered-land/data-center contracting, favorable open-book exposure and continued per-share accretion. |
One thesis, two expressions
VST and TLN should not be read as diversification from one another.
They are two versions of the same underlying trade: existing dispatchable generation inside a regional system that needs more firm capacity than it currently has.
The principal risks are correlated. A policy regime that suppresses capacity economics beyond 2029/2030, an unexpectedly fast supply response, a major reduction in real load growth or a durable collapse in forward power prices would affect both.
The difference is how that risk reaches the equity.
VST · more defensive
Heavier hedging, broader geographic exposure, large contracted relationships and a substantial ongoing buyback program.
TLN · more convex
Greater PJM concentration, a more open 2028 hedge book, significant capital-return capacity and more sensitivity to continued scarcity and new data-center contracting.
The financial opportunity is therefore not “buy two power stocks.” It is one factor expressed at two different levels of protection and upside sensitivity.
Why the equipment companies still matter
The equipment layer is not the opposing side of this thesis.
It is part of the evidence for it.
Turbine backlogs, transformer lead times, switchgear orders and transmission spending all confirm that the physical system is constrained. Companies such as GE Vernova, Eaton, Quanta, nVent and Vertiv may continue to grow strongly.
The reason MAi moved toward VST and TLN is not that hardware demand is weak. It is that investors have already awarded many of those companies premium scarcity multiples.
The generation layer is different. It owns the scarce output itself while still carrying a valuation discount for the possibility that policymakers will confiscate part of the scarcity rent.
Our conclusion is that enough of that political risk is already reflected in the stocks to create a positive expected-value setup.
What changes the conclusion
This is a dated thesis, not a permanent one.
PJM/FERC Reliability Backstop Procurement
Approval, economics, procurement size and realistic commercial-operation timing.
Policy beyond the current collar
Any move to extend or materially change price suppression beyond the 2029/2030 delivery year.
VST Q3 and FY2027 guidance
Especially the size and quality of the 2027 earnings/cash-flow bridge and the effect of ERCOT forwards.
TLN Q3
Once officially scheduled, with particular attention to the 2027–2028 free-cash-flow trajectory, hedge book and capital return.
Large-load contracting
Additional long-duration agreements can strengthen the thesis; widespread cancellations or delayed energization would weaken it.
Physical new supply
Generation actually reaching commercial operation matters more than announcements alone.
The RBP deserves special attention. It can affect market expectations before a single new turbine starts running. But its ability to eliminate the physical shortage depends on projects actually being financed, permitted, equipped, interconnected and commissioned.
That distinction — between a policy announcement and a megawatt that can actually show up on a stressed grid — is central to the MAi view.
The physical system and the political price signal are saying different things.
The U.S. does not have one national power shortage. It has regional scarcity, and PJM is one of the places where that scarcity is already measurable.
Policy has capped part of the price signal. The system nevertheless cleared below its reliability requirement.
That leaves an unusual setup: the market is valuing incumbent generators as though additional political extraction is likely, while those same companies enter the next several years with existing plants, substantial hedging, long-term contracting and aggressive capital-return programs.
MAi's 12-month distribution assigns both VST and TLN approximately 20% probability-weighted upside from their Sept. 25 reference prices, but the paths are different.
Vistra is the primary expression because more of the near-term cash flow is protected.
Talen is the higher-convexity expression because more of its 2028 economics remain open.
The opportunity exists because the physical shortage and the political price signal are saying different things. For now, MAi believes the physical system matters more.
Reference frame
This article uses the Sept. 25, 2026 closing prices of VST $138.46 and TLN $304.42 as the scoring reference. The bear/base/bull distributions are 12-month analytical ranges, not price targets.
The evidence set was frozen on Sept. 27, 2026. At freeze:
- H.R. 9340 / Ratepayer Protection Act: verified; House-passed legislation focused on recovery of incremental costs created by large-load customers.
- PJM capacity collar extension: verified; FERC accepted the relevant tariff revisions April 28, 2026.
- PJM 2028/2029 auction: verified; cleared at the $325/MW-day cap and remained 6,831 MW below the reliability requirement.
- Reliability Backstop Procurement: proposal verified; disposition remained pending in the evidence set.
- TLN Q3 2026 earnings date: not yet officially scheduled at freeze.
Primary sources
- PJM — 2028/2029 Capacity Auction Results
- PJM — Reliability Backstop Procurement Proposal
- FERC — April 28, 2026 PJM tariff-order materials, Docket ER26-1556
- GovInfo — H.R. 9340, Ratepayer Protection Act
- Vistra — Q2 2026 Results
- Talen Energy — Q2 2026 Results
Reader note
MAi Edge Capital research presents probability-weighted analytical scenarios based on the evidence available at publication. It does not represent a guarantee of future performance. Subsequent earnings, market prices, regulation, commodity prices, contracting activity and capital allocation can materially change the distribution.
