The Nasdaq Just Fell 5.5% in a Day, and AI Spending Is the Reason

The Nasdaq Sell-Off: 5.5% in a Day

The Nasdaq Composite plunged 5.5% on Thursday, November 5, 2026, and entered correction territory after falling nearly 12% from its October record. The Nasdaq 100 fell 6.4% in the same session, wiping out approximately $2.5 trillion in market value. Semiconductor stocks were hit hardest: the Philadelphia Semiconductor Index dropped 12%, with Broadcom down 10% and Nvidia down 9%. The market recovered some ground the following day, but the message of the sell-off was clear. Investors are no longer willing to take the economics of the AI buildout on faith.

This was not a single bad earnings report. It was a reassessment. The technology sell-off began the previous Friday and intensified all week, and the Nasdaq sell-off now ranks among the sharpest corrections of the year. The trigger was a mix of things, but the thread running through them is the same question: how much money is being spent on AI infrastructure, and when does it start paying for itself?

The Spending Question Nobody Can Answer

The numbers are staggering, and they are the reason investors are nervous. Amazon's 2026 capital expenditure is about $220 billion. Meta raised its guidance from $135 billion to $145 billion. Google is spending $17 billion on a single data center in Wyoming. Oracle wants $20 billion more to fund its AI cloud expansion. Add it all up and the industry is spending sums that would have been unthinkable five years ago, on infrastructure whose returns are still mostly a promise.

The bull case is that AI is the next platform shift, and the companies that build the infrastructure now will own the next decade. The bear case is that the spending is outrunning the revenue, and that the customers renting the compute cannot afford to keep paying at these prices forever. The truth is probably somewhere in between, but markets do not trade on probably. They trade on confidence, and this week confidence wobbled.

Two other factors made it worse. The week's midterm election results added uncertainty, as investors worried that increased scrutiny of data center construction could delay projects. And S&P Global Ratings moved Meta's outlook from stable to negative, a reminder that even the biggest spenders are borrowing to fund the buildout. When the ratings agencies start frowning, the market listens.

What This Means for Big Tech

The sell-off does not mean the AI boom is over. It means the market is starting to price it like a normal investment instead of a sure thing. That is a healthy correction, even if it does not feel like one to anyone holding tech stocks. Apple's record September quarter, $114.5 billion led by iPhone, shows that the big companies still have real businesses underneath the AI spending. The question is whether those businesses can fund the buildout indefinitely, and the market just said it is not sure.

Anthropic's IPO filing tells a similar story from the other side. The filing showed a $17 billion quarter and an $895 billion compute bill, numbers that are hard to hold in your head at the same time. Read the Anthropic S-1 breakdown and the tension is obvious: enormous revenue, enormous costs, and a business model that depends on the gap between them widening rather than narrowing.

The companies that survive this correction will be the ones that can show a path from spending to returns. That means real customers, real revenue, and real margins, not just bigger models and bigger data centers. The Marvell Lumentum deal, a $115 billion bet that AI runs on light, shows that consolidation is already part of the answer. Expect more of it.

What to Watch Next

Three dates matter. First, the October Consumer Price Index, scheduled for release on November 10, will tell the market whether inflation is cooperating. Second, the next round of Big Tech earnings will show whether the spending is slowing or accelerating. Third, the policy picture: how the new Congress treats data centers and AI regulation will shape the next year of investment.

For now, the Nasdaq sell-off is a warning, not a verdict. The AI buildout is real, the spending is real, and the question of whether it pays off is still open. The market just decided to stop assuming the answer is yes.