The market's loudest message came from the stocks that were supposed to have the best story. The Nasdaq Composite fell 1.33% while technology dropped 2.47%, and the damage was concentrated in AI infrastructure and semiconductor-adjacent names where expectations had been doing a lot of work. The VIX rose 4.28%, which says investors paid a little more for protection, though the move looked more like a reassessment than a scramble.
The bigger indexes hid some of the strain. The 10-year Treasury yield finished at 4.706%, and that matters because higher yields make distant profits worth less today. Investrade wrote that data-center, semiconductor and AI shares tumbled as Treasury yields pressured rate-sensitive technology stocks. Investopedia also reported that oil rose while talks to reopen the Strait of Hormuz remained stalled, adding another cost worry for companies and households.
The cost of growth moved up
The sector split made the argument cleaner. Energy gained 1.76% and health care rose 1.6%, while technology sat at the bottom of the market. That is a rotation into businesses where cash flow is either helped by commodity prices or seen as steadier, and away from companies that need investors to keep believing in expensive future expansion.
The mechanism is simple, even if the trades can look chaotic. When yields are high and oil is firm, investors ask whether a company can fund growth without leaning too hard on debt, stock issuance or flawless execution. That question is especially sharp in AI infrastructure because the demand story is large, but the bill for factories, networking gear, power and inventory is large too. A great end market can still produce rough stock moves if margins or capacity become the next bottleneck.
Fabrinet showed how narrow the path has become, falling 19.39% after earnings as investors focused on margins and cash flow despite strong results and guidance. Credo Technology dropped 13.1% after a recent run, with no verified news catalyst surfaced. Coherent lost 12.84% even after reporting strong AI-driven results, as concerns centered on factory capacity and the spending needed to expand it. Those three moves fit the same question: can the AI supply chain turn demand into profits without surprising investors on cost?
That is why yesterday felt more revealing than a normal down day. Investors didn't simply leave every growth stock at once. HubSpot rose after beating expectations, lifting profit guidance and announcing a large repurchase plan, even though revenue projections were trimmed. The reward went to the company that could point to profitability and shareholder returns. The punishment landed on companies where the next dollar of revenue looked tied to heavier spending, thinner margins or harder execution.
A small bounce has to answer a bigger question
This morning's setup is quieter, with U.S. futures pointing slightly higher and premarket trading split close to evenly. That can steady nerves at the open, but it doesn't settle the argument from yesterday. A small recovery in futures after a hard technology selloff often says traders are willing to test prices again. It doesn't say longer-term investors have stopped demanding proof.
The early mover list fits that cautious repair attempt. Micron, Allegro MicroSystems, Fabrinet and Seagate were higher before the open, which suggests some investors are willing to revisit hardware and storage names after the selloff. HubSpot was lower before the open after its prior gain, a reminder that single-company good news can fade quickly when the broader group is being judged against a tougher standard.
Asia complicates the morning without deciding it. Sharp losses in Japan and South Korea point to pressure outside the U.S., while parts of Europe and Hong Kong were firmer. That mix leaves U.S. investors with a familiar choice: treat yesterday's tech decline as a contained cleanup, or keep cutting exposure to companies where growth requires a lot of capital and time.
I think the important change is psychological. Investors still believe AI demand exists. They are starting to separate that belief from the assumption that every supplier will capture it smoothly. In a market that has already had a strong run, that distinction can move a lot of money fast.