| Account | Shares | Price | Value | % of acct | Cost/sh | Gain $ | Gain % | Target |
|---|---|---|---|---|---|---|---|---|
| 401K | 55 | $144.60 | $7,953 | 0.32% | $130.22 | $791 | +11.0% | — |
In short: An insider buy flagged from the new Form 4 tracker, in the sector he keeps naming as the multi-year opportunity. "VST CEO actually bought shares last week. So this was an interesting insider buy… a significant purchase for about a million. So not a small purchase. President and CEO James Burke bought 270k of VST, 2,000 shares at 135. It's his eighth largest trade out of 13." It sits directly under the thesis he restates later from Musk's G20 remarks: "AI growth is now being constrained by power… I think that power will be the biggest opportunity over the next few years. There's a potential 15 gigawatt shortfall just by next year alone, with AI chip production growing 40, 50% annually, while power outside China is only growing at 10 to 20%."
Vistra generates and sells electricity in the United States. It matters to this call for a reason that has nothing to do with utilities and everything to do with artificial intelligence: the constraint on building data centres is no longer chips, it is power. Singh puts a number on it from Elon Musk's remarks at the G20 — chip production is growing 40-50% a year while electricity supply outside China grows 10-20%, implying a shortfall of roughly 15 gigawatts as soon as next year.
The specific signal this week is an insider purchase. Chief executive James Burke bought 2,000 shares at $135, about $270,000 of stock, on the open market — his eighth-largest such purchase out of thirteen. Insider buying is a weaker signal than insider selling is when it is small or ceremonial, but an open-market purchase of this size by the person who sets the company's forecasts is real information: he is choosing to increase his own exposure at today's price.
It also sits inside the newest piece of SSR's tooling. Singh says the firm has started systematically tracking Form 4 filings — the disclosure executives must make when they trade their own company's stock — and is building a screener, due in October, to sort the meaningful purchases from the routine ones.
Full passage: premium transcript (PDF).
In short: One of Wapner's four exhibits that the AI-power complex is being marked down together despite NVIDIA's guidance: "Vistra is down 7%… I think there's a little more at play than 'the stocks were just up a lot.' If you get what NVIDIA had to say, there's no reason you'd think that trade should be doing what it's doing if you didn't have the whole data center debate on the front pages."
In short: Second of the three existing-plant owners Zimbardo flags as negatively affected: the specialized data-center supply contract off an already-built Pennsylvania plant is the channel the "bring your own power" rule closes. Note the scope — the order applies to projects "whether or not they connect to the larger electric grid," so even a behind-the-meter arrangement with an existing plant is caught.
Vistra is in the same business as Talen — it owns existing generating plants and has been positioning to sell their output to data centers under special long-term contracts. The new Pennsylvania rules undercut that plan for the same reason: the state now wants data centers to arrive with their own generation rather than to soak up what is already on the system.
One detail sharpens the point. The order applies "whether or not they connect to the larger electric grid," so the usual workaround — put the data center behind the meter of an existing plant and never touch the public grid — appears to be caught as well. That closes the loophole the co-location trade was built on, at least in Pennsylvania.
In short: The named expression of his top second-half theme — the AI pivot "from chip hype to power and ROI": "some power companies like VST and others that we think will do quite well," because "they may not need to buy chips every single year, but they will need to buy power every year."
Vistra generates and sells electricity. It is his cleanest expression of the theme he thinks defines the second half of the year: AI money moving from chip hype to power and actual returns on investment.
The logic is about recurrence. A data center buys chips once every few years, but it buys power every single year it operates — so power demand is the more durable annuity in the AI build-out. He is also looking at co-location companies (which house other people's servers) for the same reason.
46:59portfolio — that they think that retail should get back into the market. It's just funny. But anyway, I think that the AI pivot from chip hype to power and ROI is something I'm focused on. I think the energy and the grid bottleneck, companies that provide battery infrastructure, power, we're looking at some co-location companies that are quite interesting, some power companies like VST and others that we think will do quite well, because if you think of a power company when it comes to AI, right? That demand will sustain because
In short: The third unregulated power producer Verrone flags as weakening alongside Constellation and Talen — the AI-power trade breaking down at the generator level even while the AI build-out contractors (Quanta, EMCOR, Caterpillar) are working.
Vistra is the third unregulated generator flagged as rolling over. Placed against the same episode's positive stance on the build-out contractors — Quanta, EMCOR, Caterpillar, Schneider, Siemens Energy — it sharpens a distinction worth carrying forward.
The companies that sell equipment and construction into the AI power build are getting paid on contracts and backlogs. The companies that own the generation are getting paid on a politically visible price. When populism runs against "big anything," the second group carries a risk the first does not.
39:57You take the utilities for example. Yes. Regulated utilities have traded poorly all year. They continue to trade poorly. I think as a group they're making new lows here. The unregulated ones, the more power producers, the CEGs and the Talons and the Vistra's have also started to weaken here.
In short: The week's pick — "We are buyers of VST at today's price (notwithstanding the aforementioned technical caveats)." The premise: "at its core, the AI story is fundamentally an energy story," and Vistra is "one of the largest competitive power generators in the United States," ~40,000MW of nuclear, gas, solar and battery in ERCOT and PJM — "equivalent to around 40 large-scale nuclear power plants" — with ~5 million retail customers and "the nation's second-largest competitive nuclear fleet." At ~$145 (range $132.66-$219.82) it is 36% below the high on a ~$58B cap. Q2 2026: adjusted EBITDA $1.767B, +30% YoY, beating $1.64B consensus; generation EBITDA $994M, +68%; YTD $3.261B vs $2.589B; 97%+ commercial availability through record heat. The $4.02B revenue miss (vs $5.46B) is "real in the GAAP sense and irrelevant in the economic sense" — unrealized mark-to-market losses on forward hedges "that will settle favorably in future periods," and "the hedge program creating the apparent miss is the same driver that gives the company near-complete 2026 earnings certainty" (~100% of 2026 generation hedged; 94% 2027, 72% 2028). Demand is evidenced, not projected: "PJM and ERCOT set all-time peak loads of… 168 and 91 gigawatts in July. That is the structural AI demand proof the thesis requires, visible in real-time load data rather than forward projections." The valuation case is the exclusions: FY26 guidance of $6.8-7.6B EBITDA / $3.925-4.725B adjusted FCF (management at or above midpoint) "explicitly excludes" Cogentrix ($4.7B, ~5,500MW, H2-2026 close, "+$470-590M annual EBITDA"), the Meta nuclear PPAs (~2,600MW PJM), the Helix Platform and nuclear PTCs — so "the 2026 guidance midpoint of $7.2 billion is the starting point, not the destination… the 2028 EBITDA trajectory… almost certainly exceeds $9 billion annualized. At a $58 billion cap, the market is pricing none of it." Helix — up to $1B committed with KKR, NVIDIA and the Kuwait Investment Authority, VST as "preferred power partner" for at-scale data-center development, supplying "the one thing none of them can build quickly: 24/7 dispatchable, carbon-friendly nuclear and gas generation" — is "the stock's long-term game-changer"; plus an AWS PPA ~3,800MW at Comanche Peak and Perry Nuclear extended to 2046. Against the bearish WSJ "Heard on the Street" piece he concedes the scarcity-rent cap but relocates the moat: "the bottleneck is more than just policy, it is the physical ability to deliver reliable megawatts… the thesis is evolving, not breaking: less about passive scarcity rents and more about who can actually deliver incremental dispatchable power." Valuation: 10.6× EV/EBITDA vs Constellation ~14.5× P/E (13.6× trailing EV/EBITDA); FCF yield "the most compelling entry-point metric" at ~9.1% on the 2026 midpoint, and ~$5.4B normalized 2028 FCF = an 11.4% yield; 20 analysts Buy, avg target ~$217 (~54% upside). Bear case = ERCOT curve softness + Helix execution + PJM IRAS uncertainty, "justifiable only if all three occur simultaneously… that threshold is too high." Technical caveat stated plainly: below the 200-day for over a year, "stabilizing in the 150 range," and "to be on the safe side, you could put in a stop at 130."
Hay's starting point is that the argument about artificial intelligence — who wins, how it gets regulated, whether it takes everyone's job — misses the only question an investor can actually answer. AI runs on electricity, an enormous and growing amount of it, and electricity has to be generated somewhere by someone. So he buys the power, not the software.
Vistra is one of America's largest independent power producers: about 40 gigawatts of generating capacity — nuclear reactors, gas plants, solar and batteries — spread across the Texas grid (ERCOT) and the big eastern grid (PJM), plus a retail electricity business with roughly five million customers. To give a sense of scale, that is the output of about 40 full-size nuclear plants. It owns the second-largest fleet of privately-run nuclear reactors in the country. "Independent" here means it sells power at market prices rather than being a regulated utility with a guaranteed return.
The stock is around $145, about 36% below its high of $219.82, and the reason for the fall is a number that Hay argues means nothing. Vistra reported quarterly revenue of $4.02 billion against an expected $5.46 billion — a large apparent miss. But a power producer sells much of its electricity years in advance at fixed prices, to lock in what it will earn. Those advance contracts get marked up or down each quarter to reflect where power prices have moved, and those paper swings flow straight through the revenue line even though no money has changed hands. That is what happened here: the "miss" is an unrealised paper loss on contracts that have not settled yet. Hay's phrase is that it is "real in the GAAP sense and irrelevant in the economic sense." And there is a neat twist — the same hedging program that produced the ugly number is what makes 2026 profits almost completely certain, because essentially all of next year's output is already sold at a known price.
What the business actually did in the quarter: cash earnings (adjusted EBITDA) of $1.767 billion, up 30% and ahead of forecasts, with the generation side up 68%. The plants ran 97% of the time through record summer heat — which is exactly when a grid is desperate for them. And the demand story is not a forecast: both grids set all-time record electricity demand in July, 168 gigawatts in PJM and 91 in Texas. Hay treats that as the proof, because it is an actual meter reading rather than a consultant's projection.
The heart of the valuation case is what the company's own forecast leaves out. Vistra's guidance for 2026 — $6.8 to $7.6 billion of cash earnings — deliberately excludes four things it has already announced: a $4.7 billion acquisition (Cogentrix, adding another 5,500 megawatts), long-term nuclear power contracts signed with Meta, tax credits for nuclear generation, and the new "Helix" venture. Add them and Hay thinks 2028 earnings are north of $9 billion a year, versus a company currently valued at about $58 billion. In his words, "the market is pricing none of it."
Helix is the piece he calls the game-changer. Vistra is putting up to $1 billion into a partnership with the private-equity firm KKR, the chipmaker NVIDIA, and Kuwait's sovereign wealth fund, to be the go-to power supplier for large new data centers. Each partner brings one thing — chips, capital, patient money — and Vistra brings the one thing that cannot be conjured up quickly: power plants that already exist, already connected to the grid, that run around the clock regardless of weather. It also has a roughly 3,800-megawatt contract with Amazon's cloud arm and 2,600 megawatts committed to Meta.
He takes on the bear case directly, including a critical Wall Street Journal piece arguing that regulators will stop existing plant owners from charging whatever scarcity allows. He concedes the point and then argues it changes the shape of the opportunity rather than removing it: building new plants, transmission lines and transformers still takes years, so the winner shifts from "owns power and charges more for it" to "owns power and is one of the few able to build more." His summary: the thesis is evolving, not breaking.
On price: Vistra is valued at 10.6 times cash earnings (on the forecast that excludes all the catalysts) against rival Constellation at 13.6 times. The measure he prefers is free-cash-flow yield — the spare cash the business throws off, as a percentage of what you pay for it. On 2028's normalised numbers that is roughly 11.4%, which is a very high return for an asset this durable. Wall Street's twenty analysts average a $217 target, about 54% above the current price.
The risks are three: Texas power prices sagging as solar and wind build out faster than demand there; the Helix venture being a new and unfamiliar business to execute; and a rule change in the PJM market that could shrink the payments Vistra receives simply for having capacity available. His answer is a probability argument — the current price only makes sense if all three go wrong at once. The honest caveat is technical: the stock has traded below its one-year average price for more than a year, which is not a healthy sign. He suggests a stop-loss at 130 (i.e. sell if it falls there) rather than pretending the chart is clean. Verdict: buy at today's price.
In short: Held power exposure into the data-center build — "Vistra Energy, which we have a position in," reporting Friday pre-market.
Full passage: premium transcript (PDF).
In short: Sechan's final trade: "it's done nothing for the year but woke up this past month. I think energy can be a trade in the second half."
Vistra is a power producer — a favored way to play surging electricity demand from AI data centers. Rob Sechan made it his final trade: the stock "has done nothing for the year but woke up this past month," and he thinks "energy can be a trade in the second half" of 2026.
In short: Link power play on data-center electricity demand — a "derivative play" with better visibility than the most-popular names.
In short: "Hot money" power play likely to "get really hammered" before it becomes a wishlist buy.
Vistra is a power producer that has become a "hot money" AI-power favorite. He thinks these crowded momentum names "get really hammered" first in any pullback — so he'd wait for that washout rather than chase it here.
19:37You're bullish on infrastructure like Eaton and GE Vernova. If the consumption spike hasn't shown up yet, are these early? Well, there are three risks — NIMBY, getting the infrastructure into the location, and power. So I think you get a really good opportunity to buy the GE Vernova situations. A lot of these are really rich right now. But once this happens, a lot of the hot-money plays like the Vistras, the Constellations are going to get really hammered, and you want to create that wish list for '26.
Nothing matches this filter.
Verbatim excerpts from the public transcripts (auto-pulled at each mention's timestamp, lightly cleaned). Timestamps link into the video; "source page" opens that commentator's full analysis at this row. Click a mention's header line to expand it (one open at a time). For personal study — not investment advice.