Anthropic has taken its developer outreach to Europe with a series of events across the United Kingdom, beginning with Code with Claude London, where the company unveiled new capabilities aimed at giving enterprises greater control over how they deploy artificial intelligence.
The updates to its Claude Agents include sandbox environments that allow organizations to run AI agents within their own infrastructure, as well as “MCP tunnels,” designed to enable secure connections to internal systems without exposing them to the public internet. The move appears to directly address growing concerns among corporate leaders about governance, data security, and the risks associated with autonomous AI systems.
The London gathering, Anthropic’s first dedicated developer event in Europe, highlighted the city’s rising influence as a global AI hub. In recent months, major players such as OpenAI and other high-profile AI ventures have announced plans to expand their footprint in London, reinforcing its position as a focal point for innovation and investment. Despite poor weather, the event drew an oversubscribed crowd of enterprise users, startup engineers, and developers eager to explore Claude’s evolving capabilities.
Among startup engineers, enthusiasm for AI coding tools remains strong, with many reporting daily reliance on systems like Claude Code. While concerns about long-term job displacement persist, developers expressed a pragmatic view, favoring familiarity with existing tools over uncertainty about future alternatives. Some attendees noted recent performance issues with Claude and voiced frustration over the company’s response, but overall sentiment toward the platform remained positive.
Enterprise adoption, however, is progressing at a more cautious pace. Organizations are still grappling with fundamental questions about which processes should be automated, how to maintain human oversight, and what level of responsibility can safely be delegated to AI systems. These challenges are particularly acute in highly regulated industries such as healthcare and banking, where compliance and risk management are paramount. Executives also require reassurance and clear frameworks before committing to large-scale AI integration.
Anthropic’s head of engineering, Fiona Fung, advised companies to begin by identifying inefficient or resource-intensive workflows and evaluating whether AI tools can meaningfully improve them. Her guidance reflects a broader industry shift toward targeted, incremental adoption rather than sweeping automation.
The event’s overarching message was clear: software engineering is entering a period of significant transformation, driven by increasingly capable AI systems. While the pace and scope of change remain uncertain, both developers and enterprises are actively negotiating how best to integrate these tools into their workflows, balancing innovation with control and caution.