Mag7 Desk Note #47

Executive Takeaway

The bounce is real but thin: breadth repaired to 4-of-7 above EMA20 with six of seven green, and no name carries volume — RVOL tops out at 0.89x. All seven remain NEUTRAL with a 31% conviction ceiling, so the edge is relative, not directional. Macro is not the constraint (VIX 15.28, HY OAS 2.72%, 10Y 4.68%); the entire cohort is still hostage to one print on 8/26.

Positioning

• NVDA is the only name with momentum and peer confirmation (+6.1% vs EMA20, +5.3pp over AMD/AVGO, 84th %ile) — and simultaneously the single point of failure for breadth.
• MSFT leads its cloud comps by 8.1pp but RSI 71.8 makes it the group's worst entry, not its best idea. Fade strength; don't chase.
• Fund relative longs from peer-confirmed laggards: AAPL -6.2pp vs MSFT/GOOGL at RSI 43.0, GOOGL -5.4pp vs META/MSFT. AMZN (in-line with WMT/COST, 0% conviction) and META/TSLA (in-line with peers) are not leadership expressions.

Key Risks

• TSLA's EMA20 reclaim is one day and one headline (Swedish strike resolution) deep at +0.7% trend and RSI 47.7 — a rounding error, not a trend.
• The IV-RV gaps (AMZN -37.1pts, MSFT -27.5) sit atop 58–62% RV20 — realized-vol artifact, not free optionality.
• Equipment 21-day tape is still broken (AMAT -7.8%, KLAC -6.7%) despite strong 5-day prints; don't read ASML/LRCX strength as confirmed AI capex.

Watch

NVDA +6.1% vs EMA20 into the 8/26 print (EPS est 2.13, rev est $93.6B, IV30 41.1% ≈ realized) — lose that premium and breadth reverts to MSFT/AMZN alone.

What Changed

• Breadth 3→4: TSLA reclaimed EMA20 (-3.0% → +0.7%); META closed to -0.1% from -2.8% and lifted 21st→27th 52w %ile.
• Headline tone flipped from legal friction to constructive capex/ownership — all [NEW]: Berkshire's Alphabet stake, the $94B SpaceX mark, robotics-driven NVDA optimism.
• COMPS and Form 4 data are now populated where they were unavailable, and they validate rather than revise the prior read.