Rapid adoption of artificial intelligence across financial services is increasing banks’ dependence on a relatively small group of model and cloud-computing providers, creating concentration and operational risks alongside expected productivity gains, Moody’s has warned.
The ratings agency expects wider use of AI eventually to reduce costs and support revenue growth across banking and other financial services, but said achieving those gains will require substantial investment and could expose institutions to common points of technological failure.
Its warning focuses on the infrastructure beneath financial-sector AI. Banks can develop proprietary applications and retain control of sensitive customer and transaction data, but many still depend on external cloud providers, foundation models, specialist software, and third-party computing capacity.
Moody’s said: “The reliance of most financial firms on a relatively small set of foundation AI model and cloud computing providers risks creating a systemic dependency.”
That concentration becomes more consequential as AI moves beyond administrative assistance into processes connected with credit assessment, claims, customer service, compliance, software development, fraud detection, and other critical functions.
A disruption at a widely used provider could affect several institutions at the same time rather than remaining an isolated technology problem within one bank. Dependence can also increase exposure to supplier pricing, contractual changes, and difficult migrations where systems, data, workflows, or staff expertise become closely tied to one platform.
UK regulators have already identified concentration in technology services as a financial-stability concern. Existing work on critical third parties reflects the growing dependence of banks and other regulated businesses on shared cloud and technology infrastructure.
AI adoption across financial services is already substantial. Financial Conduct Authority research has found that more than three-quarters of surveyed financial-services companies use AI, with applications concentrated heavily in internal and operational processes.
The scale of adoption turns resilience into a procurement and governance issue. Institutions have to establish not only whether an AI system performs accurately, but which external services it depends on, where data is processed, how failures are identified, and whether a critical function can continue if an important supplier becomes unavailable.
Generative and agentic systems add further complexity. Traditional software generally executes defined processes, while more autonomous AI can retrieve information, select tools, and carry out sequences of actions. That increases the range of possible failure modes and can make technical dependencies harder to map.
Financial institutions already operate under extensive requirements covering outsourcing, operational resilience, cyber security, consumer treatment, and management accountability. AI introduces another layer through which those existing risks can emerge.
There are also significant cost implications. Banks are investing in AI partly to automate work and increase productivity while simultaneously spending on cloud infrastructure, data architecture, cyber security, model oversight, compliance controls, and employee retraining.
Productivity improvements may therefore take time to translate into stronger margins. Competitive pressure can also mean some efficiency gains are returned to customers through lower prices or better services rather than retained entirely as profit.
Moody’s has additionally highlighted the potential for AI to alter customer behaviour. More capable financial assistants could make it easier for consumers and businesses to compare rates and move deposits between institutions, adding another source of competition for funding.
That connects a technology development with liquidity management. Digital banking already allows customers to transfer money rapidly, while automated tools could reduce the effort required to compare products and respond to changes in interest rates.
Regulators are consequently examining AI through overlapping questions of innovation, consumer protection, model governance, third-party dependency, cyber security, operational resilience, and financial stability.
The commercial incentive to adopt the technology nevertheless remains strong. Banks are using AI to accelerate software development, automate service operations, analyse information, detect fraud, and reduce manual workloads.
The balance between internal control and external capability will shape future technology strategies. Larger institutions may be able to diversify providers, negotiate stronger contractual safeguards, retain more expertise in-house, and operate alternative systems, but each measure adds complexity and cost.
Supplier selection is therefore becoming part of financial-resilience planning rather than remaining a conventional IT purchasing decision.
As AI reaches further into regulated operations, the most exposed institutions may not be those using the technology most extensively, but those least able to identify, substitute, or recover the external services on which their systems depend.
You must be logged in to post a comment.