Corporate adoption of AI is surging, but the economics are increasingly strained. Companies are pouring hundreds of billions of dollars into compute, chips and data centers—often financed with short-dated, complex debt—while unit costs for large language models remain high and model iterations grow more expensive. Bloomberg pegs 2025 data-center credit deals at $178.5 billion; the BIS notes the “Magnificent Seven” now make up 35% of the S&P 500. If AI revenues don’t catch up to capex, a tech-led correction could ripple through equities, private credit and global supply chains. The U.K.’s OBR estimates a 35% global equity drop would shave 0.6% off GDP and widen the deficit by £16 billion—manageable, but painful for already fragile economies.
Related article:
BIS Annual Economic Report 2024 – market structure, tech concentration and financial stability
