Nexalytics Tech
The artificial intelligence infrastructure boom hit its sharpest reality check yet in early June 2026. Over two trading sessions — Thursday, June 4, and Friday, June 5 — U.S.-traded semiconductor stocks lost a combined $1.3 trillion in market value, according to Reuters. The Philadelphia Semiconductor Index (SOX), which tracks 30 of the largest U.S.-listed chip companies and had closed at a record high just two days earlier, fell 10.3% on Friday alone — its steepest one-day drop since the pandemic crash of March 2020 — bringing its two-session loss to about 12%. Even after the slide, the index remained up 73% for the year. The trigger was a single earnings report from Broadcom, but what followed tested something bigger: whether chip valuations had run ahead of what even strong results could support.
A Beat That Wasn't Good Enough
Broadcom reported fiscal second-quarter results on Thursday, June 3: revenue of $22.19 billion, up about 48% year over year but still short of analysts' expectations, and semiconductor AI revenue that more than doubled from the same quarter a year earlier. The disappointment was in the guidance. Broadcom projected $16 billion in AI chip revenue for the current quarter — slightly below the roughly $17.2 billion analysts had modeled — and reiterated, rather than raised, the $100 billion fiscal-2027 AI revenue target it had set back in March. Shares had climbed about 38% year-to-date, including a 15% rally in the two weeks before earnings, which is why an in-line number read as a letdown. Hargreaves Lansdown's Matt Britzman called it "a classic case of very high expectations meeting a market that wanted perfection," adding that Broadcom "is one of the more exciting names in the AI infrastructure buildout, but it also came into results as one of the higher-risk names." Broadcom shares fell more than 14% that day alone, erasing over $315 billion from a market value of roughly $2.27 trillion — one of the largest single-day wipeouts on record for any company.
The Selloff Spreads
Thursday's reaction stayed mostly within chip stocks: Marvell fell nearly 5%, while AMD, Intel, Micron, and Qualcomm slid 1.6%–6.5%. Friday was worse and broader. Nvidia dropped about 6%, wiping out more than $300 billion in market value and pulling its valuation below $5 trillion for the first time since early May. Micron tumbled 13%, erasing roughly $150 billion. Marvell — one of the hardest-hit names on both days — gave back another 17%, and AMD lost almost 11%. Broadcom itself fell a further 7.9% on Friday, bringing its two-day loss to almost 20%. "You've had a lot of people here that were just blindly buying the dip," Dennis Dick, a proprietary trader at Triple D Trading, told Reuters. "Blindly buying the dip had been winning you money, but that ended today." Wells Fargo's Ohsung Kwon offered a similar diagnosis: "The semiconductor sector was way overbought. That's why we're seeing the sell-off. I don't think it's the end of the semiconductor bull market."
Rates, Jobs Data, and a Trillion-Dollar IPO in the Background
Friday's losses were compounded by a macro surprise: the U.S. added 172,000 jobs in May, roughly double what economists had forecast, while unemployment held at 4.3%. The stronger-than-expected report reduced the odds of near-term Fed rate cuts, pushed the 10-year Treasury yield up to 4.54%, and dragged the S&P 500 down 2.6% on the day. Richly valued, growth-priced sectors such as semiconductors tend to be the most sensitive to that kind of repricing, which is part of why chip stocks fell harder than the broader market. The same week, Elon Musk's SpaceX was finishing what would be the largest IPO on record — about $75 billion at a $1.75 trillion valuation, with investor demand reportedly running roughly twice that, according to people familiar with the matter cited by Reuters. Investors were simultaneously dumping AI-adjacent chip stocks and racing to buy one of the priciest private companies ever to go public: risk appetite for AI hadn't disappeared, it had become pickier.
Reset, Not Necessarily Rupture
The case that this was a valuation correction rather than a demand problem rests on Broadcom's own numbers: AI revenue more than doubled year over year, and CEO Hock Tan told investors the company now expects to ship more than 10 gigawatts of AI chips in 2027 and that Broadcom was "very comfortable" with its memory-chip supply commitments through 2027 despite an industry-wide memory crunch. Bernstein analyst Stacy Rasgon put it plainly: "We suspect the shares may take a pause for the next couple of quarters. But the story gets interesting again once we enter 2027. If we have to wait a quarter or two for that story to re-emerge, that's OK, we'll wait for it." By Monday, June 8, early signs of stabilization were already visible, with the semiconductor ETF, Micron, Nvidia, and Broadcom all trading higher before the opening bell.
What to Watch Next
June's rout wasn't a one-off. In the week leading into July 29, 2026, 20 of the world's most valuable chip stocks lost a further $1.3 trillion combined, according to a CNBC analysis of FactSet data — a decline Morningstar chief equity strategist Michael Field called "driven largely by sentiment rather than fundamentals." Nvidia alone lost $238 billion, while memory makers SK Hynix, Samsung Electronics, and Micron lost $176 billion, $173 billion, and $113 billion respectively; SK Hynix's stock fell even though the company had just posted record quarterly profit and revenue. Forrester's Charlie Dai attributed the move to concern that AI infrastructure spending "may be peaking faster than expected" and to rising competition. Two chip-sector drawdowns of similar size within two months suggest the reflexive "buy every dip" trade of 2024–2025 is giving way to something choppier: investors still want AI exposure, but they're pricing it more nervously.
💡 The Nexalytics Take
A $1.3 trillion, two-session wipeout in the market's most important AI-adjacent sector is a real test of how much froth had built up in chip valuations — not evidence that the AI buildout itself is faltering. Broadcom's results were genuinely strong; what broke was the assumption that every quarter had to beat already-stretched expectations by an ever-widening margin. What the market actually punished was in-line guidance on custom AI chip revenue, not demand weakness — AI revenue kept climbing across the sector. A second, similarly sized chip selloff at the end of July suggests the market has absorbed that lesson only partway: "buy the dip" hasn't disappeared, but the bar for good news keeps rising, and sentiment can now erase a trillion dollars in value in days even when the underlying business is accelerating.
