AI jitters return

AI-related market weakness reflects financing anxiety concentrated in highly leveraged parts of the ecosystem rather than a broad deterioration in AI fundamentals.

• In September, the median US stock fell nearly 5% while technology stocks rose 4%; over the following three weeks, semiconductor shares fell more than 15% and South Korea's concentrated index declined nearly 20%.
• Credit spreads have widened to their highest level since spring, while Brent rose 13% in October to $116 a barrel and the US 10-year Treasury yield fell by about 30 basis points since the start of the month.
• Neocloud companies have roughly halved from their June highs and are mostly down year to date, highlighting investor concern over leveraged AI infrastructure.
• CoreWeave carries around eight turns of debt and relies heavily on a small group of customers, led by Microsoft and OpenAI, while its assets could become obsolete quickly.
• The rotation between mega-cap technology stocks and the broader market has already occurred twice in 2026, suggesting that the latest move may not yet mark a structural inflection point.
• Hyperscalers have been more resilient in both equities and credit because strong operating cash flows and balance sheets provide protection despite increased debt issuance.

Takeaway: Investors should distinguish financing stress among highly geared AI companies from weakness in the broader AI investment case, while recognising that vulnerabilities could become more serious if funding concerns spread.