UK FCA adds Anthropic to AI sandbox amid 51% application surge
Anthropic will supply Claude models to the UK regulator’s second AI testing cohort, a partnership that provides financial institutions a safe pathway to develop generative AI tools.
Anthropic will supply its Claude artificial intelligence models to the UK Financial Conduct Authority’s second Supercharged Sandbox cohort. The partnership grants 21 accepted financial organizations access to Anthropic’s technology, including Claude Code and Claude Cowork. Participating firms plan to leverage these tools to build applications spanning agent-led payments, fraud detection, compliance automation, and internal AI governance.
The FCA received 199 applications for this second group, marking a 51% increase from the inaugural round that underscores the urgency for safe AI integration. This latest cohort features a mix of established insurers, fintechs, and charities, including Scottish Widows, Money Advice Trust, and TrueLayer. By providing a controlled environment, the regulator is attempting to balance rapid technological innovation with strict consumer protection mandates.
Anthropic’s arrival builds upon the sandbox’s existing technological foundation, which is already supported by partnerships with NVIDIA and NayaOne for computing infrastructure and development tools. Concurrently, the FCA and the Centre for Finance, Technology and Entrepreneurship have introduced a 10-week Agentic Academy. This training program is designed specifically to teach firms how to build and deploy autonomous AI agents within financial services.
The regulatory architecture is deliberately segmented to match corporate maturity levels. The Supercharged Sandbox targets firms in the early discovery phase, while the FCA’s separate AI Live Testing program supports companies ready to deploy systems in actual market conditions. For investors and executives, this creates a predictable, lower-risk pathway to scale AI tools from internal prototypes to live financial products.
Managing this technological transition within regulated boundaries remains a primary operational hurdle for the industry. The integration of a leading AI developer directly into the regulatory testing apparatus signals a maturation of the market. It also provides Anthropic a direct channel to enterprise adoption at a time when financial institutions are actively allocating capital toward generative AI.
