Rising borrowing costs signal funding risks for global AI expansion: Jefferies report
As the financial markets grapple with emerging pressures, substantial investments in artificial intelligence infrastructure are evident. Companies like Oracle and CoreWeave are facing heightened credit risk, indicating struggles with borrowing amidst rising financing costs. Investors need to keep a close eye on credit spreads to gauge the viability of AI project funding.
A new report from Jefferies warns that rising borrowing costs could pose significant funding risks for the global expansion of artificial intelligence (AI) infrastructure. The research firm's "K-shaped stresses, AI tremors, and the Houthis" report, titled "Rising borrowing costs signal funding risks for global AI expansion," examines the credit risk landscape for companies involved in building AI hardware and software.
Christopher Wood, a Jefferies analyst, advises investors to closely monitor borrowing conditions as they continue to fund the massive investments required for AI growth. The report notes that credit spreads, which indicate the risk associated with lending to companies, have surged for key AI players like Oracle and CoreWeave. Oracle's five-year credit default swap (CDS) spread jumped from 186 basis points to 288 basis points, while CoreWeave's CDS spread climbed from 749 to 1,118 basis points, both representing the highest levels since the AI capex cycle began in early 2023.
The report's findings also reveal that CoreWeave's bond yield has reached 14.8%, signaling heightened financing costs. Despite this, the report does not suggest that funding for AI projects has ceased. Instead, it posits that rising credit risk is a cautionary signal for investors. The report also predicts a moderate slowdown in earnings growth among AI-related companies, with consensus earnings growth expected to drop from around 100% in 2026 to 30% in 2027 and 22% in 2028.
Written by urgent.news from The Economic Times - Top News's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.