Anthropic, the company behind Claude AI, has been selected for the UK Financial Conduct Authority’s Supercharged Sandbox as part of its second cohort. The programme brings together 21 organizations to test AI applications in regulated financial services, spanning use cases from automated compliance to fraud detection.

Applications for the cohort surged 51% year-over-year, hitting 199 total submissions.

What the sandbox actually does

The FCA’s Supercharged Sandbox is a controlled environment where firms get access to high-performance infrastructure and direct regulatory support to build AI proofs-of-concept.

The use cases being tested in this cohort cover agent-led payments, fraud and economic crime detection, compliance automation, AI governance frameworks, and tools to support vulnerable consumers. Testing runs from July through December 2026.

The sandbox builds on infrastructure support from partners including NayaOne and NVIDIA. The FCA’s AI Lab initiative, which launched in October 2024, laid the foundation for this expansion before the Supercharged Sandbox formally opened in June 2025.

Anthropic’s role is notable because it’s not a financial services company. It’s an AI model provider. Its inclusion means the FCA wants the firms building the underlying AI technology at the table, not just the companies deploying it.

The crypto-adjacent connection

The FCA maintains separate initiatives for digital assets, including a stablecoin Regulatory Sandbox and ongoing cryptoasset regulation frameworks. Several participants in the current cohort have deep ties to the digital asset ecosystem.

TrueLayer, the open banking payments platform, facilitates crypto-related transactions and has integrations with digital asset exchanges. Sardine AI Corp specializes in fraud detection and compliance for fintech companies, including those operating in crypto. Scottish Widows, the insurance and pensions giant, rounds out the notable names in the cohort.