Photo: Markus Spiske / Pexels
Barclays targets 50% developer adoption of Claude by end of 2026
The UK bank is widening its rollout of Anthropic's AI models, with half its developers expected on Claude Code by year-end
Barclays wants half of its software developers using Anthropic’s Claude Code by the end of 2026. That is the headline target in a broader expansion of Claude across the bank, announced on October 1, 2026, through an Anthropic press release.
The goal does not stop there. Barclays expects a majority of its software engineers to be on the tool by 2027, which makes this less a pilot program and more a change in how the bank writes software.
What Barclays is actually rolling out
The stated aim is to speed up software development and improve the experience for clients across Barclays’ global operations.
Claude is not new inside Barclays, though. The bank already runs a Colleague Knowledge Assistant built on the models, which has been live since 2025.
More than 16,000 staff use that assistant. It supports colleagues who serve roughly 20 million UK retail customers, and it has handled more than 1 million searches so far.
The models are also doing quieter work in the background. In Barclays’ Global Markets division, Claude sorts and routes approximately 120,000 incoming emails every day.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
The money behind the move
The rollout sits inside a multi-year efficiency program at Barclays. The bank is targeting approximately Ā£2 billion in savings over that period, with AI initiatives contributing part of that figure.
Barclays’ Co-Chief Operating Officers, Anne Marie Darling and Craig Bright, framed the expansion around two themes. They stressed the importance of managing risk, and they pointed to a shift toward more autonomous AI capabilities in software engineering and cybersecurity.
Anthropic’s growing finance client list
For Anthropic, Barclays is another marquee name in a financial services push that has been gathering pace. The company already counts hedge funds ExodusPoint, Millennium Management and Man Group among its clients.
The timing comes with a complication. Regulatory discussions around AI safety are ongoing in the UK, including conversations about potential cybersecurity threats associated with Anthropic’s models.
What this means for banks and their engineers
The existing deployments offer some evidence that the bank can operationalize this technology. The knowledge assistant’s 16,000 users and 1 million-plus searches, and the 120,000 daily emails in Global Markets, suggest Claude is already embedded in daily operations rather than parked in a lab.