Nasdaq futures declined on June 26, 2026, as semiconductor stocks extended a selloff that had already punched the Nasdaq and S&P 500 to more than one-week lows on June 23, according to The Street. The thread running through the turbulence: mounting investor anxiety over the capital costs tied to artificial intelligence infrastructure.

The damage on June 23 was sharp and broad. The Philadelphia Semiconductor Index dropped 7.9% in a single session, per Bloomberg, with the hardest hits concentrated in names that had more than doubled year-to-date — a reminder that the steepest corrections tend to arrive where the most valuation risk has accumulated. On the NYSE that day, declining issues outnumbered advancers by a 1.31-to-1 ratio, with 120 new highs against 187 new lows, according to Reuters — breadth deterioration consistent with a sector-driven rotation rather than a systemic macro shock. The rout wasn't contained to US trading hours; a broad technology and semiconductor selloff pushed global equities lower through Asian and European sessions as well, Reuters reported.

By the June 25 overnight session, futures showed the pressure hadn't fully cleared: S&P 500 futures were off 0.16% and Nasdaq 100 futures had shed 0.66%, per CNBC. Then on June 26, stock futures slipped further, with investors still focused on AI data-centre cost concerns driving the selloff.

At the single-stock level, the dispersion told part of the story. ON Semiconductor tumbled 12% in premarket trading on June 26, while Synaptics gained 6% during the same session — the latter bucking the tape in a way that points to idiosyncratic earnings or guidance news rather than pure macro pressure.

The AI spending concern at the heart of this move deserves precise framing. The bull case for semiconductor stocks through early 2026 rested heavily on a projected surge in data-centre build-out, with hyperscalers committing to eye-catching capex figures. What the market is now repricing is the other side of that ledger: whether the revenue returns on that infrastructure will materialise fast enough, and at sufficient margin, to justify the multiples at which many chip names had been trading. When a stock doubles in under six months on AI demand optimism, it takes only a modest revision to that demand outlook — or a hint that cost overruns are eroding hyperscaler willingness to spend — to produce the kind of single-session moves the SOX logged on June 23.

It is worth separating what is confirmed from what is being priced. The verified facts here are the price moves, the breadth statistics, and the directional narrative from market coverage. The magnitude of any AI capex slowdown, and whether it reflects a genuine structural deceleration or a temporary digestion of already-committed spend, remains in the realm of analyst inference rather than reported fact. Markets often front-run a thesis that takes quarters to validate — or invalidate.

For rates-sensitive context: the Reuters headline from June 23 referenced Fed hike bets alongside the tech worries, suggesting some participants are folding a higher-for-longer rate scenario into the valuation reset. Higher discount rates compress the present value of long-duration growth earnings by definition; that mechanic amplifies any sentiment shift in a sector where free cash flow is still years out for some names. Whether that rate narrative has legs depends on incoming inflation and labour market data, not on anything in these equity moves alone.

The near-term setup, as of June 26, is a market in which the Nasdaq's year-to-date gains are being stress-tested against a single underlying question: did AI infrastructure spending get priced in too early, too aggressively, or both? The SOX declining 7.9% in a session, Nasdaq 100 futures still soft three days later, and a major semiconductor name shedding 12% in premarket all point to a market that is still working through that question rather than having answered it.