Google Assembles 'Strike Team' Led by Sergey Brin to Challenge Anthropic's Code Generation Dominance

Summary

Google is escalating its efforts to develop competitive AI coding capabilities by assembling a specialized strike team, with co-founder Sergey Brin personally overseeing the initiative. The move reflects mounting pressure from Anthropic, whose advanced code generation tools and rapidly growing revenue base have prompted Google to acknowledge that its Gemini AI models lag behind in code-writing abilities. The strike team, led by Brin and Google DeepMind CTO Koray Kavukcuoglu, is focused on developing self-improving AI systems capable of writing software by reading through files and understanding user intent—initially for internal deployment on Google's proprietary codebase.

The competitive dynamics underscore a dramatic shift in the AI landscape. While Google uses AI agents to write approximately 50% of its code, Anthropic claims to have fully automated its entire codebase with AI. Anthropic's rapid ascent—evidenced by run-rate revenues exceeding $30 billion in March 2026, up from $9 billion in 2025—has forced industry leaders to reprioritize. OpenAI recently discontinued its Sora video generation model to redirect resources toward code generation, signaling the market shift. Google's strategy centers on developing models trained on its internal code, with the implicit goal of eventually applying learnings to public-facing products to close the coding capability gap.

• Google's approach focuses on self-improving AI systems trained on proprietary code for internal use before public deployment
• Competitive pressure from Anthropic is reshaping industry priorities: OpenAI deprioritized video generation to focus on code, underscoring the strategic importance of coding capabilities