Anthropic brings Claude to the FCA sandbox — 21 firms want to play 🤖

Anthropic has joined the UK Financial Conduct Authority's Supercharged Sandbox as a technology provider for the regulator's second cohort, supplying participants with access to its Claude AI models, including Claude Code and Claude Cowork. The partnership expands the FCA's program for testing artificial intelligence applications in financial services within a controlled regulatory environment, with use cases spanning agent-led payments, fraud detection, compliance automation and AI governance.

The second cohort comprises 21 organizations, including Scottish Widows, Money Advice Trust and TrueLayer. The FCA said Wednesday that it received 199 applications for the latest round, a 51% increase over the first cohort, signaling rising demand from financial firms seeking supervised settings to experiment with AI systems. Selected participants will build applications using Claude alongside infrastructure and development tools provided through the FCA's existing partnerships with NVIDIA and NayaOne.

Alongside the cohort launch, the FCA opened a 10-week Agentic Academy run with the Centre for Finance, Technology and Entrepreneurship to help firms develop and deploy AI agents for financial services. The Supercharged Sandbox targets firms in the discovery and experimentation phase of AI development, while the FCA's separate AI Live Testing program is designed for firms ready to bring AI systems into live markets.

The initiative comes as policymakers in the United States and United Kingdom weigh clearer frameworks for digital assets and AI in finance, with market participants watching whether coordinated rulemaking can keep pace with rapid deployment of agent-based systems across trading, lending and compliance functions.