AI's $4 Trillion Wipeout Erases Nasdaq's Gains for the Year
A brutal week took roughly $4 trillion off the AI sector and left the Nasdaq flat for the year, and the damage ran the full length of the supply chain.

A massive AI correction wiped out about $4 trillion in market value across the sector and dragged the Nasdaq down another 5%, erasing all of the index's gains for the year. For executives and boards, the week is a live stress test of how much of their plan, compensation and capital structure depends on a single concentrated trade.
The number is the story: about $4 trillion in market value was wiped out across the AI sector, and the Nasdaq slid another 5% to close out one of the worst weeks in recent memory. That single week was enough to erase every gain the index had banked for the year, leaving it flat. The damage ran straight down the AI supply chain, hitting Sandisk, Micron, Seagate, Applied Digital, Broadcom, Nvidia and Dell. This was not a rotation out of one crowded name. It was a correction that moved through the entire stack at once.
Read the "flat for the year" line again, because it is the part that should get a board's attention. An index that spent months climbing gave all of it back in days. For anyone whose portfolio, compensation or hiring plan was built on the assumption that AI-linked equities only move one direction, the week was a live demonstration of how fast a concentrated trade can unwind. The Nasdaq is now exactly where it started the year, and the AI complex is where the losses are concentrated.
Here is why one theme can drag an entire index. The AI trade is not a single stock; it is a chain. Nvidia designs the accelerators. Broadcom supplies networking silicon and custom chips. Micron, Sandisk and Seagate sit in memory and storage, the commodity end of the stack, where pricing has historically been the most cyclical. Dell builds the servers that house it all. Applied Digital sits in the data center capacity layer. When the market decides the buildout is worth less than it thought, it does not reprice one link. It reprices the chain, because every link's revenue depends on the same capital expenditure budget.
A correction of this size is not the same thing as a collapse in the underlying business. Historically, sharp drawdowns in a sector that has run hard tend to reflect positioning and expectations more than a sudden change in demand. Deadline's report does not attribute the week's damage to a single catalyst. What it does document is scale: roughly $4 trillion of market value gone, and a benchmark index that gave back a full year of gains in a handful of sessions.
The second-order effect is where this gets expensive. Data center buildouts are funded with a mix of cash flow, debt and equity. When the equity value of the companies doing the building falls, the cost of that capital typically rises, and projects with the longest payback periods are the first to get re-underwritten. That is the mechanism by which a bad week in the market can turn into a slower year for the companies selling into it. Memory and storage names like Micron, Sandisk and Seagate are the most exposed to that shift, because their pricing has historically swung harder than the rest of the stack.
The pain is not evenly distributed. Employees paid heavily in stock at AI-adjacent companies watch their compensation reset in real time. Retail investors who bought the theme near the top are now underwater. Index funds tracking the Nasdaq are flat for the year, which means the passive investor who thought they were diversified was, in practice, making a concentrated bet on the same handful of names. And founders and executives with concentrated positions face the classic problem: the wealth is on paper, and paper moves.
Watch the memory and storage complex first. Micron, Sandisk and Seagate are the cyclical edge of the AI trade, and they typically signal whether this is a sentiment reset or something deeper. Watch Broadcom and Nvidia, the two names most associated with the AI narrative. Watch Dell, where server demand is the clearest read on whether enterprise buyers are still spending. And watch the index itself: a Nasdaq that is flat for the year is a Nasdaq with no cushion, which means every subsequent data point lands on a market with nothing left to absorb it.
For executives and boards, the lesson is not to sell everything. It is that concentration cuts both ways, and the same theme that produced the gains can produce the drawdown. The practical work is unglamorous: stress-test the plan against a market that values your sector materially lower, check how much of your compensation structure and treasury sits in correlated assets, and make sure the capital allocation decisions you approved in a bull market still clear the bar in this one. The AI buildout is not over. But the market just repriced what it is willing to pay for it, and that repricing is now a planning assumption, not a hypothetical.
This story's Key Insights and Take-aways are locked.
Create a free account to unlock Executive Actions for one credit.
Register to UnlockAlways free for Executives Club members. Join the Club
More in Technology
Anthropic mathematician claims Claude helped crack the Riemann hypothesis
If the proof holds, AI has toppled math's most famous open problem; if it does not, it is a case study in announcing breakthroughs before anyone verifies them.
Anthropic mathematician says AI proved the Riemann hypothesis. Verification will take time.
A single unverified claim from inside Anthropic now sits against more than 160 years of open mathematics, and the gap between asserting a proof and surviving peer review is the entire story.
Anthropic mathematician claims proof of Riemann hypothesis, math's 167-year holy grail
If the AI-generated proof survives peer review, it would settle the problem mathematicians have chased since 1859 and cap a year in which the technology ran the table on math.


