AI data centers need enormous amounts of electricity, and the companies that supply that power are a second-order way to play the AI boom without paying inflated chip-stock prices. Vistra sold off recently, yet its CEO put $1.17 million of his own money i

Verdict: Real demand, thin evidence — wait for the triggers, not the narrative

The demand thesis is visible in the filings: Monolithic Power's June 2026 quarter revenue rose 21.9% quarter over quarter to $980.6M with net margin at 26.2%, and Vistra's fiscal 2025 free cash flow of $1.3B and 18.5% return on equity sit in the top decile of utilities — per the September 5 Yahoo Finance pieces, the AI power-delivery story shows up in reported numbers, not just rhetoric. But the thesis's most quoted fact cuts against it: over the disclosure window covering the period ended June 30, 2026, Monolithic Power insiders were net open-market sellers of roughly $99.8M and Vistra insiders net sellers of about $9.3M, so the celebrated $1.17M CEO purchase is one data point against roughly $109M of aggregate selling. Vistra's latest quarter is the weak link — revenue down 12.0% to $4.40B, net income down 70.4% to $305M, and debt-to-equity up to 3.57. The mechanical evidence rests on a single completed 60-month trade that returned 19.3% but endured a 12.4% drawdown, with exits approximated on daily bars. Neither entry condition set is live today — for VST, the ADX (14) above 20 requirement is the binding blocker at 8.7; for MPWR, price sits $114.90 below its $1,338.76 EMA trigger — so the disciplined action is to wait for the mechanical triggers rather than pre-position on the story. Parameter-sensitivity work produced no recommended setup, meaning no robust nearby-parameter configuration was established; the published thesis-consistent triggers are what you are relying on.

Trade now: neither leg is triggered — here is exactly what you are waiting for

Neither entry condition set is live today, so the correct action is to wait rather than buy the thesis early. On the VST leg, the stock closed at $149.30 — just $1.96 above its 50-day EMA of $147.34 (needs a cross above; close). RSI (14) at 70.0 is already above the 45 threshold, and price is $7.06 above the lower Bollinger band at $142.24 (needs price at or below that band; close). The binding blocker is ADX (14) at 8.7 versus the required reading above 20 — about 11.3 points away. On the MPWR leg, the stock closed at $1,223.86: price sits $114.90 below its 50-day EMA of $1,338.76 and RSI (14) is at 31.9 versus the 45 threshold, while the below-Bollinger condition (band at $1,313) and the ADX condition (64.6, above 20) are already met — two of four conditions live. If a leg triggers, the risk framework is mechanical: positions are sized to a fixed 2.4% risk per trade, the hard stop hits at a 2.4% unrealized loss, and the profit target is a 4.8% unrealized gain — roughly 2:1 reward-to-risk — with positions capped at 25% of capital and a 60-bar time exit as the backstop. The 60-month backtest on this rule set shows one completed trade returning 19.3% with a 12.4% maximum drawdown; note that exits were filled on daily bars, so the reported drawdown and win rate are coarse. Parameter-sensitivity work produced no recommended nearby-parameter setup, so we are running the thesis-consistent triggers as published. Concretely, "wait" means: for VST, watch for a price cross above $147.34 alongside ADX (14) rising above 20 while price holds at or below $142.24 — the ADX requirement is the one that will take real time. For MPWR, watch for price reclaiming $1,338.76 and RSI (14) crossing back above 45; with RSI at 31.9 and price 27.6% off its range high, that leg is further from trigger, but the oversold conditions are already in place. Do not pre-position — the 2.4% stop discipline only works if the entry is mechanical.

The power-chokepoint thesis has receipts: a CEO buying the dip and accelerating power-chip revenue

The core of this idea — that electricity and power delivery are the choke points of the AI buildout — is visible in the latest numbers from both names. Monolithic Power's June 2026 quarter shows revenue of $980.6M, up 21.9% from the prior quarter, with net margin expanding to 26.2% and operating margin to 31.0%. Even before that, MPWR's FY2025 revenue grew 20.8% year-over-year, placing it in the top third of its IT sector peer group, and the cited Yahoo Finance piece (September 5, 2026) headlines a 48% revenue jump — the demand acceleration the thesis leans on is real in the filings, not just the narrative. The Vistra half of the argument is behavioral. Per the Yahoo Finance article, CEO James Burke bought $1.17 million of stock after the selloff — a signal that leadership views the dip as overdone. That purchase sits alongside fundamentals that are strong at the full-year level: Vistra's FY2025 revenue of $17.6B grew 19.1% year-over-year, free cash flow of $1.3B puts it in the 91st percentile of utilities, and return on equity of 18.5% is in the 90th percentile of its sector. This is not a distressed company being bought by a desperate insider; it is a cash-generative utility whose own CEO is adding on weakness. The…

Scores

• Conviction score breakdown: 45
• Thesis support: 72
• Trade readiness: 30
• Risk quality: 42
• Backtest evidence: 25
• Fundamentals trend: 58

Watch items

• VST — ADX (14)
• VST — Close vs 50-day EMA
• VST — Close vs lower Bollinger (20)
• MPWR — RSI (14)
• MPWR — Close vs 50-day EMA
• MPWR — Insider net open-market value
• VST — Insider net open-market value
• MPWR — Quarterly revenue (SEC filing)
• VST — Quarterly net income (SEC filing)
• VST — Debt-to-equity
• MPWR — Dividend ex-date
• VST — Dividend ex-date