UK's HSBC Warns Against AI-Fueled Overreach in Global Credit Markets

HSBC’s research team cautions that current market excitement around artificial intelligence risks outpacing the technology’s near‑term benefits for corporate credit, creating a vulnerability to abrupt repricings. While AI can lift productivity and margins for a subset of firms, HSBC emphasizes the uneven timing and industry concentration of those gains; many borrowers will see little immediate credit improvement. Market pricing has begun to reward perceived AI winners, compressing spreads and weakening compensation for default and liquidity risk — a dynamic that leaves portfolios highly sensitive to disappointment. External market evidence reinforces HSBC’s concerns: industry estimates point to roughly $3 trillion of AI‑focused data‑center projects under consideration and potential hyperscaler procurement of about $1.5 trillion, concentrating demand and delivery risk in a few large players. Stress work from other banks, notably UBS, models how such concentration and capex‑driven shocks could lift defaults materially in leveraged, illiquid private‑credit pools, with a severe scenario implying cumulative defaults well above historical norms. Practitioners at major managers are already shortening effective holding periods, tightening covenants and formalizing rapid operational playbooks — steps that can both mitigate and crystallize repricing across similar borrowers. Public credit desks and bond investors have begun to penalize smaller software vendors and narrowly focused suppliers, widening spreads where margin sustainability or compute‑funding pathways are unclear. The buildout of specialized compute also concentrates upstream hardware and foundry demand, reshaping supplier allocations (for example, reallocations toward high‑performance DRAM and HBM) and increasing the chance of delivery timing shocks that alter borrower cash flows in a single investment cycle. HSBC advises investors to shift from theme‑driven allocation toward granular credit analysis: focus on balance‑sheet resilience, earnings quality and realistic scenario testing, and design hedging and liquidity plans that allow for sharper repricing. For investment‑grade credits the bank recommends prioritizing issuers with durable free cash flow and strong covenant protection; in high‑yield, selectivity is crucial, particularly for cyclical borrowers whose AI upside is speculative. On a market‑structure level, compressed spreads reduce liquidity cushions and can magnify volatility if expectations shift; regulators and risk managers should therefore monitor leverage, covenant quality, redemption terms and short‑term funding that could amplify corrections. In short, HSBC reframes AI as an accelerant of dispersion and timing risk — a structural force that demands disciplined credit selection, active management and stronger contingency planning rather than passive exposure to a popular narrative.