A broad selloff in chip and AI-linked stocks reflected investor anxiety that the trillion-dollar capex boom may not generate commensurate returns.

Meta's disclosed forward commitments and capex outlook erased tens of billions in market value and dragged data center suppliers like GE Vernova, Vertiv and Eat

A completed round at that valuation would make OpenAI the most highly valued private company ever.

Tesla's capex reversal intensified the debate over whether AI infrastructure spending is peaking, weighing on chip demand expectations.

Growing local opposition to data centers is emerging as a political risk for AI stocks heading into the US midterm elections.

The policy change comes weeks after an OpenAI engineer said cutting models off from the internet would not work.

Broadcom's Hock Tan said the next AI bottleneck is not chips, power or memory, and analysts remain bullish despite financing questions.

Bank of America's forecast underscores how large the AI infrastructure buildout could grow, and the financing it will require.

Rising yields and concerns over debt-funded AI buildouts are pressuring neocloud and hyperscaler-related names.

Prominent investors are split on whether AI valuations are stretched, with the IMF and academics also weighing in on bubble risk.

A supply warning from a semiconductor equipment maker compounded fears about slowing AI-driven chip demand, hitting Marvell, AMD, Intel, Micron and Western Digi

Reuters reports Anthropic is pursuing a listing while other AI companies delay their IPOs amid market turbulence.

The round is one of China's largest, coming just after Washington ordered federal agencies to remove DeepSeek technology.

BofA and Morgan Stanley both published lists of AI infrastructure stocks positioned to withstand a capex slowdown.

The unit is seeking capital for what would be South Korea's largest AI data center bet.

Analysts see Google's TPU program as a growing competitive threat to Nvidia's AI accelerator dominance.

TI's outlook signaled AI demand is spilling into analog semiconductors.

TSMC and ASML signals point to a multi-year AI chip demand cycle.