Semiconductor stocks suffered their worst single day in months on June 23, 2026, dragging the broader Nasdaq down with them. The Philadelphia Semiconductor Index fell 7.9% that day alone, per Bloomberg. By June 26, the pressure hadn't let up: Nasdaq futures were still sliding, and investors remained spooked by one underlying worry—whether the money spent building AI infrastructure will actually pay off.

The damage was broad and fast. On June 23, declining stocks outnumbered those that gained by roughly 1.3 to 1 on the NYSE, according to Reuters, with 187 new lows against only 120 new highs. This pattern—losses spreading across the market rather than hitting a single stock or sector—pointed to a sector-wide rotation rather than panic over the whole economy. The selloff wasn't limited to US markets either; technology stocks fell in Asian and European trading as well, Reuters reported.

By June 25 overnight trading, S&P 500 futures were off 0.16% and Nasdaq 100 futures had fallen 0.66%, per CNBC. The following day brought fresh declines. At the individual stock level, though, the picture grew messier: ON Semiconductor tumbled 12% in early trading on June 26, while Synaptics gained 6% in the same window—a divergence that pointed to company-specific news rather than a broad sector collapse.

Here's what matters about the underlying worry. Through early 2026, chip stocks rode high on excitement over data-centre construction. Tech giants like Amazon, Google, and Microsoft publicly committed to spending vast sums—tens of billions a year—on the servers and infrastructure needed to run AI models. The market priced in the assumption that all that spending would happen, and happen fast.

Now investors are asking a harder question: will that infrastructure actually generate enough revenue, and at fat enough profit margins, to be worth what was spent? When a stock has doubled in six months on AI excitement, even a modest downward revision to that story—or news that costs are eating into profits—can trigger sharp one-day drops like the 7.9% plunge on June 23.

Here's the key distinction: the price moves, the breadth numbers, and the direction of trading are facts. What is not yet confirmed is the scale of any slowdown in AI spending. Whether the pullback is temporary—companies just pausing between waves of spending—or signals a deeper problem in the economics remains analysis and inference, not settled fact. Markets routinely jump ahead of outcomes that take months or quarters to play out.

The Fed's interest rate path adds another layer of pressure. Higher rates make future profits worth less in today's money—a basic valuation mechanic called "discounting." For semiconductor companies that won't generate strong cash flow for years, higher rates compress valuations faster. The June 23 Reuters headline tied the selloff to both AI concerns and bets on the Fed holding rates elevated, suggesting some investors are now pricing in that combination: slower AI spending plus the weight of higher financing costs.

As of June 26, the market is still working through a core question: did AI infrastructure spending get priced into chip stocks too early, too aggressively, or both? The Nasdaq's year-to-date gains are under real stress, and the selloff shows no sign of resolving quickly.