Anthropic Hosts First European Developer Event Showcasing Claude's AI Coding Capabilities

Anthropic showcased its AI coding tool, Claude Code, at its first European developer event in London, where it's gaining dominance among startups. While developers embrace features like its 7-layer architecture for complex tasks, co-founder Jack Clark warns AI poses existential risks and could achieve "recursive self-improvement" by 2028. The company is also changing its pricing model, moving programmatic use to a metered credit system starting June 15, 2026.

Claude Code Event and Capabilities

• ▪Claude Code is built on a structured, seven-layer architecture designed to handle large codebases with persistent, layered context.
• ▪Anthropic hosted its first European developer-focused event, "Code w/ Claude," in London starting on May 19.
• ▪Claude Opus 4.7, released in April 2026, features a 1 million token context window, significantly larger than competitors like Grok 4.
• ▪A feature in Claude Managed Agents called "dreaming" allows coding agents to record notes and learn about a specific code base over time.

Developer and Startup Adoption

• ▪Anthropic claims that most of the software at the company, including in Claude Code itself, is now written by its own AI.
• ▪At the "Code with Claude" event, a large number of developers admitted to shipping code written by Claude without reading it first.
• ▪Companies including Spotify and Delivery Hero have reshaped their software development teams around Claude Code.
• ▪A Business Insider survey of startup founders and VCs found a growing consensus that Claude Code is becoming the default AI coding tool.

AI Safety and Risk Mitigation

• ▪Anthropic silently patched a SOCKS5 hostname null-byte injection vulnerability that allowed for a sandbox bypass in Claude Code.
• ▪Anthropic developed a model, Claude Mythos, with nation-state-level cyber-offensive capabilities that it decided not to release publicly.
• ▪Anthropic co-founder Jack Clark warned that AI poses a "non-zero" chance of killing everyone and could reach "recursive self-improvement" by 2028.

Pricing and Business Model Changes

• ▪The billing change affects the Claude Agent SDK, non-interactive CLI mode, and third-party agents, while interactive use remains unchanged.
• ▪Under the new model, subscribers receive a monthly credit ($20 for Pro, $100 for Max 5x, $200 for Max 20x) that does not roll over.
• ▪Effective June 15, 2026, programmatic use of Claude will move to a separate monthly credit pool billed at full API rates.

Claude Code Technical Features

• ▪Developers can create custom "skills" to extend Claude Code with project-specific tools and spawn "sub-agents" to work on large tasks in parallel.
• ▪Claude Code's architecture uses `CLAUDE.md` files for persistent context and supports event-driven "hooks" for programmatic guardrails.

Competitive Landscape

• ▪Compared to xAI's Grok, Claude is considered superior for deep reasoning and writing, while Grok leads in real-time social data access.
• ▪Startup founders describe Microsoft's GitHub Copilot as no longer providing meaningful advantages and view Cursor as a "fading secondary tool."
• ▪GitHub's Copilot is also moving to usage-based billing on June 1, 2026, nine days before Anthropic's change.