AVGO AI Miss Triggers 300B Wipeout|Dow Record Masks Chip Repricing

The Rotation Signal

Broadcom posted the strongest AI chip quarter in its history on Wednesday evening and lost $300 billion in market value by Thursday afternoon. Revenue of $22.2 billion grew 48% year-over-year, AI semiconductor revenue surged 143% to $10.8 billion, and the company guided third-quarter revenue to $29.4 billion — beating consensus. The stock still fell 14.8%.

The gap between the reported numbers and the market reaction is not a misread. It is the repricing of a positioning structure. Broadcom had added roughly $270 billion in market capitalization across the five sessions leading into earnings, with institutional accumulation that the Chaikin Money Flow showed fading from a reading of 0.5 in early May down to 0.11 by the day of the print. That divergence — price rising while institutional buy pressure leaked — is what made the Q3 AI chip guidance of $16 billion a structural trigger rather than a minor miss. The buyside estimate had been $17.2 billion, and Broadcom did not raise its full-year AI revenue forecast of $100 billion.

What the selloff clarified is which capital class was carrying the position into earnings. Retail and momentum flows had been the marginal buyer through the five-session pre-earnings rally, while institutional hands had been quietly distributing. When the forward number came in below the whisper, the momentum structure unwound and the price gave back gains faster than fundamentals alone would justify.

The sector contagion followed the same logic. Micron fell 7.7%, AMD dropped 4%, Intel slid 3% — none had company-specific news. The move was not about their individual outlooks; it was about unwinding a correlated trade that had been built on the premise that every AI infrastructure name would print upside surprises sequentially. Broadcom broke that sequential expectation, and the capital that had been rotating through the sector hit an exit simultaneously.

What the rotation did not resolve is whether the repricing represents a pause or a ceiling. CEO Hock Tan disclosed on the call that Google would likely draw on multiple chip suppliers — not exclusively Broadcom. That single sentence shifted the competitive read on Broadcom's custom ASIC moat, which is distinct from the guidance miss itself and was not fully priced into the after-hours reaction.

The Consumer Side

The same session that saw Broadcom's AI chip positioning unwind delivered a separate repricing on the consumer side that did not share the same capital pool but deepened the day's overall signal. Lululemon cut its full-year revenue forecast to $11 billion to $11.15 billion — a decline of 0% to 1% versus prior year — despite posting a Q1 beat of $2.5 billion in revenue, up 4%.

The guidance cut is the surface number. What sits underneath it is more specific and more durable. Interim co-CEO Meghan Frank named two causes directly: disappointing product launches and negative media commentary. The media commentary stemmed from the company's unresolved proxy battle with founder Chip Wilson, whose public campaign to overhaul the board disrupted brand messaging across the Americas and China through the first quarter. In North America, comparable sales fell 5% on a constant-currency basis. In China — a market Lululemon had counted as a growth counterweight — the controversy amplified.

Foreign institutional net flows and retail positioning in Lululemon had been built around the international recovery thesis, which held that a 13% international comparable sales gain would offset North American softness. That thesis is still mathematically intact in Q1. But the revised full-year guidance implies management does not believe the international buffer holds through the year. The capital that had been holding Lululemon on the international-growth premise now faces a revaluation of the time horizon attached to that premise.

The incoming CEO, Heidi O'Neill from Nike, does not join until September. That gap is not a neutral interval. It is the period during which a brand under active public criticism operates without a permanent executive who can claim the credibility to reset the narrative. The positioning pressure is therefore not symmetrically distributed across time — it is concentrated in the next three months, before O'Neill arrives and before the Q2 print gives investors the first read on whether the guidance cut was conservative or directional.

The connection to the day's broader theme is not sector correlation but repricing mechanism. In chips, the question was whether a guidance number below a whisper estimate warranted a $300 billion drawdown. In consumer, the question is whether a brand under founder conflict produces lasting demand destruction or recoverable noise. Both questions share the same underlying structure: is the forward pressure a flush or a trend?

The Verification Point

The Dow's 874-point gain on Thursday — its largest in months and a fresh record close — needs to be read precisely for what it was and what it was not. Healthcare stocks and banks led the advance, not because new fundamental catalysts emerged in those sectors, but because capital rotating out of chip and extended tech positions needed a same-session destination. Goldman Sachs gained 5%, Fifth Third Bancorp rose 4.7%. The Dow's composition, which is price-weighted and tilted toward healthcare and industrials, made it the mechanical recipient of the rotation. The S&P 500 gained only 0.4%. The Nasdaq fell 0.1%.

Iranian ceasefire signals contributed to the session through oil's 2.8% decline to $95.03 per barrel for Brent, which compressed bond yields slightly and gave rate-sensitive smaller-cap names room to move. The Russell 2000 gained 1.4%. But the oil move was a supporting condition, not the driver — the rotation was already underway before the ceasefire headlines accelerated it.

The monitoring variable coming out of this session is narrow. Broadcom's $16 billion Q3 AI chip guidance represents 200%-plus growth year-over-year. The number is structurally large. What the market priced on Thursday was not that AI infrastructure spending is slowing, but that the rate of upside surprise relative to a pre-earnings positioning structure had peaked. Whether that repricing holds through the next session depends on whether institutional buyers re-enter Broadcom on the dip — and there is direct evidence to watch. Bank of America and Morningstar both raised their price targets to $650 and $600 respectively following the earnings, and at least 22 analysts lifted targets to a median of $500, implying more than 4% upside from the Thursday close.

The reading that Thursday's selloff was a positioning flush rather than a cycle ceiling holds as long as Broadcom's next quarterly AI revenue print — expected around September — comes in above the $16 billion Q3 guidance. The reading breaks if institutional net flows into Broadcom remain negative through the next four to six weeks while the buy-side AI capex estimate for 2027 is revised down, which would indicate the competitive read from Hock Tan's Google supplier comment is being priced beyond the single-session reaction.