Moody's: Banks Racing Into AI Risk Becoming Dependent on a Few Tech Giants
A new Moody's report says banks' rush to adopt AI could leave them exposed to outages, price hikes, and job cuts tied to a handful of Silicon Valley model providers.
What happened: Credit rating agency Moody's has warned that banks and insurers piling into AI are becoming increasingly dependent on a small cluster of Silicon Valley firms that supply the underlying models and cloud computing. Its report says this could leave the financial sector exposed to outages, cybersecurity gaps, fraud, and pricing power wielded by tech firms under pressure to turn a profit. More than 75% of UK financial firms already use AI, per a Treasury select committee report, mostly to automate admin work, process insurance claims, and assess creditworthiness.
Why it matters: Moody's says AI will eventually cut costs and boost revenue for banks, but heavy upfront investment and fierce competition mean much of that payoff gets 'competed away.' Meanwhile relying on a few foundation-model and cloud providers creates what it calls 'systemic dependency' — a single outage could spread across many banks and customers at once. It also flagged 'vendor dependence risk': providers like OpenAI and Anthropic, both still losing money, may eventually push up prices for the AI services banks now depend on.
How it works, plainly: Most banks don't build their own AI models — they license access to a handful of foundation models and cloud platforms, then plug in their own data. That concentration is efficient but fragile: if one provider goes down, raises prices, or changes terms, many banks feel it simultaneously. Banks say they retain control over their most valuable asset, proprietary customer data, and some are hedging with open-source models, multiple vendors, and tough contract negotiation to avoid getting locked in.
The rollout: Lloyds Banking Group is pressing ahead anyway, committing £13bn to an AI strategy that includes £2bn of cost cuts. CEO Charlie Nunn acknowledged the plan will 'impact work' and require reskilling and new hiring. Moody's separately estimates roughly a one-in-five chance AI can match a capable mid-level employee's output by 2030. It also warned AI could make it easier for depositors to shift large sums between banks quickly, straining trust in weaker institutions, and expects regulators to scrutinize this concentration more closely.