Broadcom's CEO says AI demand could be 'significantly' higher than what's being built, but land and power are the choke points — meaning electricity suppliers and grid-builders now hold the scarce resource. Nvidia guiding 70% growth and Amazon committing

Verdict: The AI-power thesis earns a wait, not a buy — the entries aren't armed and the insiders are selling

The thesis is real and visible in the numbers: Nvidia guided roughly 70% fiscal 2028 revenue growth (per the September 4 Yahoo Finance piece) and Amazon committed to 2 million GPUs, so grid power and electrical equipment are the bottleneck, and Constellation just printed $11.1B in Q1 2026 revenue with a 26.9% gross margin while Quanta's Q2 revenue grew 21% and its operating margin jumped to 7.3%. The strongest point against is the insider tape: as of the June 30, 2026 ownership cycle, disclosed filings show net open-market selling of roughly $123.2M at Quanta, $9.3M at Vistra, and about $785K at Eaton — the people closest to the supply chain are reducing exposure just as the thesis peaks. The second problem is mechanical: none of the four names has armed the entry yet — Eaton (RSI 48.2, ADX 50.7) and Quanta (RSI 47.3, ADX 53.5) are closest, needing only a few points of RSI cooling, while Constellation (RSI 73, ADX 13.7) and Vistra (ADX 8.7) likely need weeks and a pullback. The completed CEG backtest returned 27.1% over 24 months on just two trades with a 12.9% max drawdown — directionally consistent but thin, and the upcoming September 30 quarter reports are the real fundamental catalyst that either validates pricing power or hands you the entry on better terms. A confirmed RSI-45 print on Eaton or Quanta with the EMA structure intact, or an earnings guide-up on grid/data-center backlog, is the fact that would flip this verdict from wait to buy.

Trade now

**Nothing is triggered yet — and that is the honest read.** The strategy wants each name to satisfy four conditions at once: price below its 21-day EMA, RSI (14) at or below 45, the 21-day EMA having crossed above the 50-day EMA, and ADX (14) above 20. As of the latest close (CEG $299, ETN $411, PWR $624, VST $149), only Eaton and Quanta are close: both trade below their 21-day EMAs and show strong trend readings (ADX of 50.7 and 53.5 respectively), with RSI just above the line at 48.2 and 47.3 — a few points of cooling each would complete the picture. Constellation is the furthest away, with RSI at 73 and ADX at 13.7; Vistra's RSI of 69.9 and ADX of 8.7 put it in the same camp. 'Wait' here means literally standing aside until a name prints all four conditions on a daily close — no partial entries. **Risk is defined before entry, which is the point.** Once a position is opened, the rules cap loss at a 2.4% stop on unrealized value and take profit at 4.8%, with an additional exit at a 127.2% Fibonacci extension and a time stop after 60 trading days. That is an effective reward-to-risk of roughly 2:1 per trade. The completed backtest on the CEG pair returned 27.1% over 24 months across 2 trades with a 12.9% maximum drawdown — supportive of the setup, though on only two completed trades, so treat it as directionally consistent rather than statistically thick evidence.…

Scores

• Conviction score breakdown: 61
• Thesis support: 78
• Trade readiness: 40
• Risk quality: 60
• Backtest evidence: 55
• Fundamentals trend: 72

Watch items

• ETN — RSI (14)
• PWR — RSI (14)
• CEG — RSI (14)
• CEG — ADX (14)
• VST — RSI (14)
• ETN — EMA (21) vs EMA (50) cross
• PWR — EMA (21) vs EMA (50) cross