A rare divergence has opened between Big Tech (the AI capex spenders) and semiconductor stocks, with semis up 80%+ this year while Magnificent Seven names sit in correction territory. The gap historically resolves through mean reversion — either semis give back gains as capex expectations reset, or Big Tech recovers and validates the semi premium.
A rare divergence has opened between Big Tech (the AI capex spenders) and semiconductor stocks, with semis up 80%+ this year while Magnificent Seven names sit in correction territory.
NVDA, AMD, and AVGO are priced for sustained hyperscaler AI capex while the Magnificent Seven sit in correction — the question is whether semis are leading the market higher or are the last shoe to drop.
If Magnificent Seven names report strong AI capex guidance next earnings cycle, the Mag7 recovery closes the gap by moving Big Tech up rather than semis down — the short semi leg gets squeezed into strong fundamentals.
CoverageSource: MarketWatch · Published here MON, JUN 29 · 4:08 PM ET · the only report in this recordHow this is decided →
A notable spread has emerged between semiconductor stocks and the 'Magnificent Seven' hyperscalers that buy their chips. Semis have surged more than 80% this year on AI demand euphoria, while the very companies writing the capex checks — Alphabet, Meta, Microsoft, Amazon — have slipped into correction territory, creating a rare and historically unstable divergence.
The enrichment data underscores why semis have been bid so aggressively: NVDA is running $215.9B in revenue with 71.1% gross margins and 55.6% net margins — extraordinary economics that justify a premium. AMD has grown revenue 34% YoY to $34.6B, though net margins at 12.5% are far thinner. AVGO at $63.9B revenue and 67.8% gross margins is also a high-quality compounder. The fundamental case for these businesses is real.
The tension, however, is structural: if the Magnificent Seven are correcting because investors are questioning the ROI on AI capex, that same skepticism is a direct headwind to the chips powering those workloads. Semis are a high-beta derivative of hyperscaler spending intentions — a capex slowdown or reprioritization would hit NVDA and AMD order books before it hits revenue, and sentiment could reprice faster than fundamentals.
The bull case is that Big Tech's correction is a valuation reset in high-multiple names, not a signal of reduced AI spending — in which case semi earnings power (especially NVDA's 55% net margin) remains intact and the gap closes upward. The bear case is that the divergence is a leading indicator: Big Tech is smarter money about its own capex outlook, and semis at 80% YTD gains are priced for perfection heading into a potential spending pause.
What to watch: any Magnificent Seven guidance cut on AI infrastructure spend, NVDA's next earnings print and data center segment growth, and whether the SOX index starts to roll over relative to the Nasdaq. The divergence is the signal — the direction of resolution is the trade.
The 80%+ YTD semi rally is priced off hyperscaler capex intentions, but the Magnificent Seven correcting while semis hold highs is a historically unstable spread. NVDA's 55.6% net margin and AMD's thin 12.5% net margin suggest NVDA is far better insulated, but both face the same demand-signal risk if Big Tech guidance disappoints. A pairs short semis / long Mag7 captures mean reversion without a pure directional bet on AI spend.
The read above, as written. kept as written · closes shown from JUN 30 on
4-8 weeks. Follow to be told when one lands.
Price context does not establish that the story caused the move.
NVDA's 71.1% gross margin and 55.6% net margin — best-in-class among large-cap semis — suggest the earnings power justifying elevated semi valuations is real and the Magnificent Seven correction is a valuation, not a spending, story.
Semis are a high-beta derivative of hyperscaler capex commitments, and the Magnificent Seven correcting into their own earnings cycle historically precedes a semi de-rating as forward spend assumptions get trimmed.
Kept as written · your side, if you take one, is graded privately against licensed closes after 10 trading days · nothing here is advice · How the Wire is made →
Only names the read names · 3M line, licensed closes · no proxy basket.
NVDA +2.63% since the story · 1 trading day · −2.27% over 3 sessions
Stories on NVDA: the first close moved a median −0.22%, up 72 of 151.
Full record →Reaction = the first close after the story against the close before it. Prior-session closes only; not a call.
Shares a name with this story — discovery, not a connection.
This page is kept as it was written on Jun 29. Later coverage joins it only when the company and catalyst evidence match, and what the stock did is shown from licensed end-of-day closes — never re-graded, never backdated. The judgment is yours.