Barclays is targeting 50% Claude Code adoption among its developers by the end of this year

Barclays and Anthropic said Thursday they are expanding their partnership, with the British bank setting a specific, dated target: 50% of its software developers using Claude Code by the end of 2026, with Anthropic's Claude models reaching the majority of Barclays engineers by some point in 2027. The expansion also covers cybersecurity work and continues two deployments already running inside the bank -- a client-support assistant used by thousands of UK staff, and an email-processing system in Barclays' Global Markets division.

"Claude now helps 16,000 Barclays colleagues find answers for customers, sorts 120,000 emails a day," Anthropic's chief commercial officer Paul Smith said, calling the expansion "a significant milestone for us in the U.K." Barclays framed it around engineering capacity rather than headcount: Group Co-Chief Operating Officer Craig Bright said "software engineering and cyber security are both being reshaped by increasingly capable AI systems," while his co-COO, Anne Marie Darling, tied the rollout to customer and colleague outcomes rather than cost. Neither company published a contract value, a seat count, or a per-user cost -- the only hard numbers on the table are usage figures, not spend.

Barclays isn't breaking new ground so much as catching up to a pattern already set at scale. Anthropic has Claude in production at JPMorgan, Goldman Sachs, Citi, AIG, and Visa; Goldman alone pairs roughly 12,000 engineers with Devin- and Claude-based coding agents and has pushed its own AI assistant to more than 46,000 employees. A PYMNTS Intelligence survey cited alongside this announcement found 95% of banks and insurers now broadly deploying AI tools in back-office functions -- which makes Barclays' announcement less a signal that banks are adopting AI, a settled question by now, and more a data point on how specific the public targets are getting. "50% by year-end" is a number a bank can be held to in a way "we're exploring AI" never was.

It also lands two days after a comparable enterprise-AI story on this desk: Meta's new Enterprise Platform, Anthropic's own Claude Marketplace, and Microsoft's Work IQ are now competing directly for the same corporate budgets that deals like this one draw from. Barclays' announcement names Anthropic specifically rather than a marketplace or platform -- a direct enterprise relationship, not a procurement-through-marketplace purchase -- which is itself a data point on how the largest, most risk-sensitive buyers are still choosing to transact.