After one of the most extraordinary rallies in semiconductor history — a 71% surge across a single quarter fueled by collective faith in artificial intelligence's transformative promise — the market paused to ask whether ambition had outrun reality. On the first trading day of Q3 2026, chip stocks from Micron to Intel to AMD shed billions in market value, not because the AI story had ended, but because a single report about Meta's excess computing capacity forced investors to confront a quieter, more unsettling question: what happens when the infrastructure of the future arrives before the fut
Chip stocks tumble as AI rally cools; Micron drops 11% after record quarter
The question now was whether the industry had built too much, too fast
Mark
So the chip stocks had just finished their best quarter ever, and then they immediately fell apart. What changed overnight?
Mimi
Meta announced it might rent out excess computing capacity. That single report made investors wonder if the whole industry had overbuilt — if demand wasn't keeping pace with all the infrastructure being installed.
Mark
But Meta's stock went up. How does that work?
Mimi
Meta benefited because investors saw it as smart — the company could now monetize its investments. But for the chip and equipment makers, the implication was that maybe the buildout was slowing.
Luke
We should be careful here. The report was about Meta's plans, not confirmed demand destruction. One company's excess capacity doesn't necessarily mean the whole sector overbuilt.
Mimi
True. But it was enough to spook investors who had just ridden a 71% rally in two months. They were looking for a reason to take profits anyway.
Mark
So this is just profit-taking after a huge run, not a fundamental shift in AI demand?
Luke
Probably both. The profit-taking was real — Micron down 11%, Intel down 9%. But the Meta news gave investors permission to question whether valuations made sense anymore.
Mimi
And that's the real story. After the rally, investors started asking harder questions about whether all this infrastructure spending would actually turn into earnings.
Mark
Do the analysts think the sector is still worth owning?
Mimi
Some do. Treasury Partners' chief investment officer said he still favors the big tech companies investing in AI, because their earnings growth is still strong. But the market is clearly more skeptical now.
Luke
The key phrase in the reporting is 'looking for clearer evidence.' Nobody knows yet if the investments will pay off. That uncertainty is what's driving the volatility.
Le Pouls
- A 71% quarterly surge in semiconductor stocks collapsed into a 5%-plus single-day drop, erasing confidence as swiftly as it had been built.
- Meta's reported plan to rent out surplus AI computing capacity sent a shockwave through the sector — if the world's biggest AI spender has more than it needs, the entire buildout may be ahead of actual demand.
- Chip equipment makers like Lam Research, KLA Corp., and Applied Materials — companies that had more than doubled in Q2 — fell at least 10%, suggesting the selloff cut deeper than surface-level profit-taking.
- Meta itself rose over 9%, rewarded for its ability to monetize excess capacity, even as the suppliers who built that capacity were punished for the same excess.
- Some investors, like Treasury Partners' Richard Saperstein, held their ground, pointing to strong earnings fundamentals — but the market's mood had shifted from euphoria to scrutiny.
After one of the most extraordinary rallies in semiconductor history — a 71% surge across a single quarter fueled by collective faith in artificial intelligence's transformative promise — the market paused to ask whether ambition had outrun reality. On the first trading day of Q3 2026, chip stocks from Micron to Intel to AMD shed billions in market value, not because the AI story had ended, but because a single report about Meta's excess computing capacity forced investors to confront a quieter, more unsettling question: what happens when the infrastructure of the future arrives before the future does. It is the oldest tension in technological history — the gap between what we build and what we are yet ready to use.
The semiconductor sector entered the third quarter in retreat, with the VanEck Semiconductor ETF falling more than 5% on Thursday — a jarring reversal from the day before, when it had just closed out its strongest quarter on record. Between April and June, the index had surged 71% as investors poured money into the chipmakers expected to power the AI boom, broadening their bets well beyond Nvidia to include memory specialists and processor manufacturers.
Micron fell 11%, Intel slid 9%, and AMD dropped 7% — three companies that had collectively added nearly $2 trillion in market value during Q2 alone. The selling extended to semiconductor equipment makers, with Lam Research, KLA Corp., and Applied Materials each declining at least 10% after having more than doubled in the prior quarter.
The catalyst was a report that Meta Platforms was considering renting out excess AI computing capacity it had already built. For investors, the implication was pointed: if one of the world's largest AI infrastructure spenders had more computing power than it could use, perhaps the entire industry had overbuilt. Meta's stock rose more than 9% on the news — analysts saw opportunity in its ability to monetize surplus capacity and expand into enterprise AI — but for its suppliers, the signal was sobering.
Not all investors fled. Some pointed to continued earnings strength among major technology companies as reason to stay constructive. But the day's trading made clear that the AI rally had entered a new phase — one defined less by collective enthusiasm and more by a harder question: whether an industry that built so much, so fast, had built more than the world yet knows how to use.
The semiconductor sector opened the third quarter on shaky ground Thursday, with investors rushing for the exits after one of the most explosive rallies in recent memory. The VanEck Semiconductor ETF, which tracks the industry's major players, dropped more than 5% — a sharp reversal from the previous day, when it had just closed out its strongest quarter on record. Between April and June, the index had surged 71% as money poured into chip makers expected to power the artificial intelligence boom.
Micron Technology, the memory chip manufacturer that had been among the biggest beneficiaries of that wave, fell 11%. Intel slid 9%. Advanced Micro Devices declined 7%. Together, these three companies had added nearly $2 trillion in market value during the second quarter alone, as investors broadened their AI bets well beyond Nvidia and began betting on rising demand for the memory chips and processors that would be needed to support the infrastructure buildout. That confidence evaporated in a single trading session.
The selling pressure extended beyond the chip makers themselves. Semiconductor equipment manufacturers — the companies that build the machines used to manufacture chips — took even steeper hits. Lam Research, KLA Corp., and Applied Materials, all of which had more than doubled during the second quarter, each fell at least 10%.
The trigger for the reversal was a report that Meta Platforms was considering renting out excess artificial intelligence computing capacity it had built. The news sparked a fundamental reassessment among investors: if Meta, one of the world's largest spenders on AI infrastructure, had more computing power than it needed, perhaps the entire industry had overbuilt. Perhaps demand was not keeping pace with the frantic expansion of data centers and hardware. Perhaps the valuations that had driven the rally were no longer justified.
Meta's own stock moved in the opposite direction, rising more than 9%, as investors viewed the company's ability to monetize its infrastructure investments as a positive development. Analysts at KeyBanc Capital Markets suggested the move could help Meta expand into the enterprise AI market and generate faster returns from its billions in annual infrastructure spending. But for the companies that supply the chips and equipment that go into those data centers, the implication was darker: the gold rush might be slowing.
Not everyone had abandoned the sector. Richard Saperstein, chief investment officer at Treasury Partners, said he remained constructive on the large technology companies investing heavily in AI, arguing that their earnings growth remained strong even as valuations had moderated in response to concerns about heavy capital expenditure. The sharp reversal, though, highlighted a new reality in AI-related stocks after an extraordinary run — investors were becoming more selective, looking for concrete evidence that the massive sums being poured into infrastructure would actually translate into sustainable profits. The question that had driven the rally — whether AI would transform computing — remained unanswered. But the question that now mattered more was whether the industry had already built too much, too fast.
Citations marquantes
Earnings growth remains strong even as valuations have moderated due to concerns over heavy capital expenditure— Richard Saperstein, Chief Investment Officer, Treasury Partners
The move could help Meta expand into the enterprise AI market and generate quicker returns from its infrastructure investments— KeyBanc Capital Markets analysts
Envie de l'histoire complète ?
Lire l'original sur The Economic Times ↗
Technology