The number that detonated the semiconductor market last week was not a loss. Broadcom reported record revenue of $22.19 billion for its fiscal second quarter, up 48 percent from a year earlier, and AI chip revenue grew 143 percent, with bookings for the quarter above $30 billion against $10.8 billion shipped. The company beat revenue estimates, raised its full-year AI target earlier in the year, and reiterated a fiscal 2027 goal of more than $100 billion in AI chip revenue. Wall Street sold it anyway.
Broadcom shares fell 13.4 percent in the session after the report and another 7.9 percent on Friday, a two-day decline of roughly 20 percent. The trigger was the forward number. Broadcom guided third-quarter AI chip revenue to about $16 billion, against a consensus of $17.2 billion, a shortfall of roughly $1.2 billion, and chief executive Hock Tan declined to raise the full-year 2026 AI revenue target of $56 billion, even as first-half AI revenue approached $19 billion. Some investors had been pricing in an upgrade, and when it did not come, a beat was treated like a miss.
Two comments on the earnings call deepened the selloff. Tan acknowledged that Google would likely draw on multiple chip suppliers, a hint that Broadcom’s flagship custom-silicon customer may be spreading its orders, and he warned that the rapid growth of AI chip sales was diluting the company’s overall gross margins. “Semiconductor margins remain very stable and very solid. It’s the mix,” he said, pointing to the shift toward lower-margin AI chips.
Friday added a macro shock to the sector-specific disappointment. The May jobs report came in substantially stronger than expected, raising the probability that the Federal Reserve would hold rates higher for longer, a poor backdrop for stocks valued on years of future earnings. The Philadelphia Semiconductor Index fell 10.3 percent, its worst session since March 2020, erasing about $1.3 trillion in market value from U.S. chipmakers. The Nasdaq Composite dropped 4.18 percent to close at 25,709.43, its largest single-day decline in more than a year, and the Dow fell roughly 680 points.
The damage was broad, but the pattern was telling. Marvell, a Broadcom rival in custom silicon, fell 16.7 percent; Micron dropped 13.2 percent, losing about $150 billion in market value; Arm Holdings fell 12.8 percent; Intel 11.3 percent; AMD 10.9 percent. Even Nvidia, whose business has shown no signs of slowing, fell 6 percent, losing more than $300 billion in a day. Traders noted that Micron fell on nothing specific to its own business, evidence that the AI trade has become a single correlated position: when one data point in the chain disappoints, the whole complex reprices together.
The analysts split into two camps. Some framed the pullback as a buying opportunity, arguing that demand remains intact and that the selloff punished valuation rather than fundamentals. Stacy Rasgon of Bernstein, who tracks the custom-silicon market closely, noted the shortfall against what investors had modeled and said a quarter or two of patience would likely re-emerge as the story strengthens. Ohsung Kwon, chief equity strategist at Wells Fargo, said the sector had been way overbought but stopped short of calling it a structural break. “I don’t think it’s the end of the semiconductor bull market,” he told Reuters.
The data supports the demand side of the argument. The Semiconductor Industry Association reported April global chip sales up 93.9 percent year over year, and Broadcom’s own bookings of more than $30 billion against $10.8 billion shipped show orders far outpacing shipments. The sector had risen about 75 percent year to date before Friday, which cuts both ways: it explains the violent correction, and it shows how much optimism was already in the price.
The episode has also refocused attention on Broadcom’s software arm, built around its $69 billion acquisition of VMware, which generates the high-margin subscription revenue that cushions the lower-margin AI chip business. Investors said the selloff may have been amplified by positioning: with the sector up roughly 75 percent this year, much of the AI trade was owned by momentum funds that exit on any sign of deceleration. The buying opportunity argument rests on whether the next round of hyperscaler capital spending numbers, due in the coming weeks, confirms that demand is broadening rather than slowing.
The deeper question is whether the selloff reflects a pause or a repricing. Tan’s comment about Google spreading orders points to a structural shift, as hyperscalers diversify custom chip supply and AI capital spending rotates from concentrated to distributed. For a market that had priced continuous acceleration, any sign of diversification reads as deceleration. Bulls argue that is precisely the sign of a healthy cycle, with demand spreading beyond a handful of vendors. The coming weeks will show whether buyers step in at these levels, or whether the market needs to see the acceleration itself, not just the promises, before it pays up again.
