Friday, 30 October was the day the AI trade stopped being priced on ambition. The Nasdaq Composite slid 2.9% — its largest one-day percentage drop since 5 June — and the two catalysts were earnings: Intel’s guidance and Amazon’s spending forecast. Reuters described the close as a rout driven by AI cash-burn fears; Bloomberg framed it as Big Tech’s $750 billion AI bet colliding with slowing core growth.
Intel’s Worst Day Since the 1970s
Intel shares collapsed 29% after the chipmaker guided to adjusted EPS of 25 cents on revenue of $15–16 billion, against consensus of $17.1 billion and 43 cents. Two things compounded it: an explosion at an Israeli fab that will cost roughly $600 million in lost revenue over two quarters and halted 10nm shipments, and a PC slowdown management now expects to run “at least through the first half of next year.” Counterpoint data shows global PC shipments down 4% in Q2. D.A. Davidson downgraded to neutral with the note “Not fixed yet.” The US government’s 10% stake lost about $12 billion in value; Nvidia’s stake, taken under last September’s agreement, shed roughly $6 billion.
Amazon fell 1.4% after a Q3 miss on both lines while holding its 2026 capex forecast at $220 billion — its second consecutive quarter of negative free cash flow. Alphabet, which already dropped more than 12% on Thursday, fell again; Microsoft had trimmed its own forecast to $170 billion from $175 billion and Alphabet held $195–205 billion while reporting more than $9 billion of negative free cash flow. Nvidia lost 4%, Broadcom 5%, Qualcomm 11%.
The Neocloud Contagion
The damage spread furthest to compute-only providers: CoreWeave, Nebius, IREN and Oracle all fell as the market read the hyperscaler numbers as a demand signal. This is the tail of the October 19 Gartner shock — a forecast that 80% of neoclouds could cease operating as independent companies, with more than $3.5 trillion in AI infrastructure spending potentially exposed to impairment by 2030. Chief forecaster John Lovelock’s framing was that the frontier-model market looks like table salt sprinkled into existing businesses, with an eventual Anthropic-OpenAI duopoly.
The read for builders: the constraint is no longer willingness to spend, but whether it earns. For agents and inference, the same arithmetic that rattles markets — utilization, revenue per GPU-hour, negative FCF — is now the budget conversation.
Related: Wall Street closes sharply lower as AI cash burn fears spark tech selloff, Intel stock has worst day since 1970s, AI buildout hits snag as Google and Microsoft put off raising capex forecasts