A broad selloff in AI-linked shares erased trillions in market value in October as investors questioned the durability of hyperscaler capital spending.

Intel's guidance miss tied to weak PC demand and rising memory costs dragged chip stocks lower and wiped roughly $12 billion from the US government's stake.

Amazon reported soaring spending alongside slower online retail growth, with its CEO flagging AI costs as a barrier to broader adoption.

Alphabet's plan to keep lifting capital expenditure disappointed investors even as Google's in-house chips lifted cloud revenue.

Samsung said an AI-driven memory shortage will worsen and kept investors waiting on details of returns from its foundry and memory investments.

Off-balance-sheet and contracted spending commitments are under growing scrutiny as hyperscalers commit record sums to AI infrastructure.

Analysts and investors are increasingly questioning whether the debt-financed AI buildout can be sustained as bond markets push back on capex.

S&P's negative outlook revision signals credit raters are beginning to weigh AI capital spending against cash generation at large platforms.

A junk rating would raise Anthropic's borrowing costs as it funds large compute and infrastructure commitments.

Broadcom's large custom-chip deal with Anthropic is seen as a potential catalyst even as chip peers sold off in the wider AI trade.

Jassy said TSMC capacity constraints and customer demand pushed AWS to consider selling its custom silicon beyond its own cloud.

MediaTek said AI demand growth will continue 'for sure' and expanded its AI ASIC production with a new customer.

The arrangement would see the chipmaker backstop financing for a customer's data center, a sign of deepening vendor exposure to AI buildout debt.

Memory makers fell even as some analysts argued strong AI demand and contracts make the selloff a buying opportunity.

Insurers and lenders are exploring ways to transfer risk tied to the growing debt financing of AI data centers.

Neocloud and AI infrastructure names sold off sharply on concerns that hyperscaler spending growth may be peaking.

The Chinese AI chip designer is growing sales while still losing money, reflecting the competitive domestic accelerator market.

The research firm flagged the scale of spending that could be exposed if AI infrastructure economics fail to hold up.