A ratings agency looked at how banks buy AI and reached for the words 'systemic dependency'

In a report dated 10 August, Moody's warned that the banking sector's rush into AI has left it leaning on a very small group of suppliers, and named OpenAI and Anthropic directly. Both are loss-making and under investor pressure to reach profitability — pressure Moody's argues could later be converted into leverage over the pricing terms of the institutions built on top of them. Because so many financial firms depend on the same narrow set of foundation model and cloud providers, the agency says an outage at one could propagate quickly across customers and sectors. It also describes a "circular AI ecosystem": hyperscalers report multibillion-dollar backlogs, much of it contracted to pre-IPO AI labs they have themselves invested billions into, and those labs spend the money back on their investors' cloud. The figures are not small. More than 75% of UK financial services firms already use AI, heaviest among insurers and international banks, mostly for admin automation, claims processing and credit assessment; Lloyds Banking Group alone has committed £13 billion to an AI strategy that includes £2 billion of cost cuts. Moody's puts roughly a one-in-five chance on AI matching the output of a mid-level employee by 2030.

✓ Verified · 3 sources

Related stories