A Jefferies research note points to mounting pressure on capital markets to keep pace with the worldwide build-out of artificial intelligence infrastructure, with debt-financing conditions for companies at the centre of that expansion becoming less favourable. Credit-risk gauges tied to Oracle and CoreWeave moved higher according to the document, feeding into broader questions about whether capital markets can sustain the scale of spending that AI build-out now demands.
In his GREED & fear newsletter, under a section headed "K-shaped stresses, AI tremors and the Houthis," Christopher Wood pointed to lending and borrowing conditions as a watchpoint for investors, noting that corporate spending on artificial intelligence infrastructure had been accumulating at a rapid rate. According to the document, those holding positions in the market are advised to watch credit spreads closely, as any widening could indicate that appetite for financing the continuing build-out of AI infrastructure is beginning to wane.
When companies borrow, the spread between their debt yields and benchmark rates reflects how much additional return lenders demand to take on that exposure. Widening spreads tend to reflect a market environment where lenders and buyers of debt require higher returns to offset the perceived likelihood of default.
The report's data show that Oracle's five-year credit default swap spread climbed from 186 basis points on 11 September to 288 basis points, while CoreWeave's equivalent measure moved from 749 basis points up to 1,118 basis points across that same window. A CoreWeave bond set to mature in 2032 was trading at a yield of 14.8 per cent, a level that points to the company carrying a heavier-than-average burden on its debt obligations.
According to the report, spending at this point in the AI capital expenditure cycle — which began taking shape in early 2023 — has not previously reached these levels. Building out AI infrastructure calls for heavy capital outlays across data centres, computing hardware, and the broader technology stack that supports it. Elevated borrowing rates tend to put debt-financed ventures under strain, particularly where businesses draw heavily on outside capital to get them off the ground.
Nothing in the document points to a halt in funding directed at artificial intelligence initiatives. Rising credit risk, the report notes, functions as an advance indicator worth tracking by those with market exposure. The prospect of decelerating profit growth at AI-linked firms was among the concerns Jefferies drew attention to.
Drawn from a separate publication titled Diversifying Beyond AI — released on 13 October — the data pointed to a projected deceleration in consensus earnings growth for global AI stocks, from roughly 100 per cent in 2026 down to 30 per cent in 2027 and further to 22 per cent by 2028. According to the document, AI-linked equities carried a forward price-to-earnings multiple of roughly 23, nearly double the 13-times multiple attached to non-AI counterparts — a spread that points to a pronounced divergence in how the two groups are valued by the market.
While AI spending remains at elevated levels, mounting concerns over financing costs and the prospect of moderating profit growth are placing firms under growing scrutiny to demonstrate the returns on their technology outlays.
