Pulse24 Original
Chip Stocks Lost $1.4 Trillion in a Day, Their Worst Drop Since the Covid Crash. Friday Clawed Back a Third of It.
November 7, 2026
Semiconductor stocks had their worst single session since the 2020 Covid crash on Thursday, erasing roughly $1.4 trillion in combined value before clawing back about a third of it on Friday. Nvidia's November 17 earnings report is now shaping up as the market's next verdict on AI spending.
The PHLX Semiconductor Index dropped 1,257.52 points on Thursday, a 12.25% decline that left it at 9,011.14. That is the index's largest single-session loss since March 16, 2020, when the Covid crash took it down 15.90% in one session. A Reuters markets columnist described Thursday's move as the second-biggest one-day decline on record for the index, and the semiconductor-tracking ETFs told the same story: SOXX fell 12.15% to $131.90, its worst day since March 2020, and SMH lost 11.21% to close at $447.73.
The damage spread well beyond chipmakers. Micron closed down 17.46% at $656.39. AMD fell roughly 17%. Intel slid about 10%. Nvidia, the sector's largest company, held up comparatively well, down 9.26% to $181.60. The Nasdaq Composite fell 5.52%, its biggest daily decline in 19 months. By one estimate, chip and AI-linked names lost close to $1.45 trillion in market value on Thursday alone, with Nvidia's share of that loss running near $446.6 billion. A broader estimate put the damage across the entire S&P 500 at roughly $3 trillion for the day.
[[IMG1]]
Why Chip Stocks, and Why Now
None of the explanations traders circulated required a single bad piece of chip news. Start with Tuesday's election results. Democrats won control of Congress, and investors spent Thursday pricing in what that could mean for AI data center construction, including a wave of state-level opposition to new facilities and the prospect of tighter federal rules on the industry. CoreWeave had already given markets an early read on that reaction when it fell 9.5% the day after the vote, in what was an election story more than an earnings one.
A separate explanation has less to do with politics and more to do with crowded trades. Nomura's Charlie McElligott pointed to hedge funds unwinding positions in this year's biggest winners to lock in gains before year-end, a flow-driven move rather than a reassessment of chip fundamentals. Nationwide's Mark Hackett offered yet another angle, citing weak results from two chip designers and a wave of institutional selling as funds trimmed their borrowed positions.
Equipment and Memory Fell Hardest
Look at which stocks fell the most, and a pattern emerges. Lam Research, which makes the equipment chipmakers use to build next-generation chips, fell 16.36%. Micron, which sells memory into AI servers, fell even harder at 17.46%. Nvidia, which sells the chips that actually run AI workloads today, fell less than both at 9.26%.
That ordering matters. Equipment and memory sit further up the supply chain, tied to spending decisions that get made years before a chip ships. When those names fall harder than the company selling finished AI processors, it suggests investors are pricing in a slowdown in future capital spending commitments, not a drop in current demand. It is the same tension that showed up when Microsoft's AI spending came in below forecast while Alphabet's capex jumped and its stock still fell: the market has grown less forgiving about capex, even when the underlying growth numbers hold up.
Friday's Partial Rebound
Chip stocks clawed back some of the damage on Friday, though not all of it. The semiconductor index closed up 4.21% at 9,390.25. AMD led the rebound, up 11.04% to $449.04. Micron added 5.12% to close at $690.00. Intel rose 7.71%, and Nvidia gained 2.78% to close at $186.64. SOXX and SMH rose 3.51% and 3.63%, respectively. In dollar terms, the sector recovered roughly $469 billion of Thursday's $1.45 trillion loss, around a third of what was wiped out.
This was not the first rough week for chip stocks this earnings season either. The semiconductor sector fell 3.8% on a Monday in October after Oracle's debt filing and Marvell's acquisition spree raised separate questions about how AI infrastructure is being financed. Thursday's drop was far larger, but the underlying question investors keep returning to is the same one: whether the capital being committed to AI buildouts can be justified by the revenue that eventually shows up. The Nasdaq-100's own measure of that unease came a day earlier, when it confirmed its third correction of 2026, a reminder that this year's AI trade has given back sizable gains more than once.
What to Watch Next
Nvidia's fiscal third-quarter report, scheduled for after the close on November 17, is the next real catalyst. The company has not confirmed the date publicly, but it is the one set by Wall Street's earnings calendars, and it will carry more weight than usual. Investors will be watching for the same signal they were looking for in AMD's own recent results, where data center revenue nearly doubled but guidance pointed to an even faster quarter ahead. Any sign that Nvidia's own customers are pulling back on orders, rather than just adjusting the pace of future spending, would confirm the harsher read on Thursday's selloff. A clean beat with confident guidance would support the gentler one: that this was a liquidity event, not a demand one.
Also worth tracking: how state and federal lawmakers follow through on the data center opposition that traders cited as a catalyst, and whether hedge fund positioning data in the coming weeks shows the kind of broad de-grossing that McElligott described. Both will say more about whether Thursday was a one-off flush or the start of a more lasting repricing of AI infrastructure stocks.
The Pulse24 Take
A 12% drop in a semiconductor index is the kind of move that invites a simple story, and Thursday got two competing ones. One says the AI trade is cracking under its own weight, that spending commitments built on optimistic demand forecasts are starting to look fragile next to an increasingly skeptical Washington. The other says this was mechanical: a handful of hedge funds trimming winning positions into year-end, amplified by thin November liquidity, with the political framing serving mostly as a convenient trigger rather than a real cause.
Both stories can be partly true at once. What tends to separate a healthy pullback from the start of something worse is whether the names that fell hardest, the equipment makers and memory suppliers sitting earliest in the AI supply chain, keep underperforming once the initial panic fades. Friday's rebound was broad enough to suggest dip buyers still believe in the trade, but it recovered only a third of what was lost, which is not the kind of snapback that ends a debate. Nvidia's November 17 earnings call will not settle whether AI infrastructure spending is sustainable over the next five years. It will say a great deal about whether the market still believes it is, for now.
How we read the data
Curious how we get from raw data to a take like this? Our Trader's Toolkit walks through the tools we lean on.
Explore the Toolkit