FCA brings Anthropic on board for second Supercharged Sandbox cohort
FCA brings Anthropic on board for second Supercharged Sandbox cohort #
The UK’s Financial Conduct Authority has brought Anthropic on board as a technology partner for the second cohort of its Supercharged Sandbox. Participating firms will have direct access to Claude-family AI models as they develop and test financial services applications in a regulated environment.
Under the arrangement, the 21 organisations selected for the programme can use Claude Code and Claude Cowork, Anthropic’s developer-focused and collaborative AI tools, alongside existing infrastructure provided by NayaOne and NVIDIA’s enterprise AI computing platform. The sandbox is designed to let firms move from concept to tested prototype under regulatory supervision, with the FCA observing and engaging throughout the process.
The second cohort drew considerably more interest than the first: the regulator received 199 applications, a 51 per cent increase on the 132 received when the programme launched in 2025. Among the 21 firms selected are Scottish Widows, Money Advice Trust and payments infrastructure provider TrueLayer. Participants will focus on a broad set of AI use cases, the FCA’s announcement said, including agent-led payment services, fraud and financial crime detection, AI governance tooling, improved access to financial services for underserved and vulnerable consumers, and compliance automation.
Jessica Rusu, the FCA’s chief data, intelligence and information officer, said the high volume of applications reflected strong industry appetite for environments where experimentation can take place safely and responsibly, and that Anthropic’s involvement would give firms the technical capability to move faster.
Alongside the cohort announcement, the regulator revealed the launch of the Agentic Academy, a ten-week specialist programme run in partnership with the Centre for Finance, Technology and Entrepreneurship. The academy is intended to deepen participating firms’ capability in designing and deploying agentic AI systems. The FCA has identified agentic applications as a priority for cohort two, with a focus on autonomous AI agents used for tasks such as customer service, compliance monitoring, and payment execution.
The Supercharged Sandbox is part of the FCA’s broader AI Lab initiative, which also includes a separate AI Live Testing programme. The regulator has said it will publish an evaluation report covering the second cohort’s outcomes in the first quarter of next year, following a showcase event later in 2026.
Copy link