Google has reportedly formed a specialized internal group known as a “Strike Team” aimed at improving the coding capabilities of its artificial intelligence models, particularly within its Gemini ecosystem. The initiative comes amid increasing competition in the AI development tools space, where rivals such as Anthropic and OpenAI have made significant advances in automated coding systems. According to reporting, the move reflects growing internal acknowledgment within Google DeepMind that Anthropic’s coding focused tools currently outperform its own models in several areas of software development.
The formation of the Strike Team follows internal discussions around performance gaps between competing AI systems. Anthropic has positioned its Claude based tools as highly effective for coding tasks, with claims that a significant portion of its codebase is now generated with AI assistance. In contrast, internal estimates suggest that Google’s current AI assisted coding rate is around 50 percent. This disparity has become a key motivation behind the effort to strengthen Google’s capabilities in this specific domain. The focus on coding performance is also linked to broader strategic goals, including advancing AI systems toward higher levels of autonomy and potential self improvement, an idea reportedly supported by Google co founder Sergey Brin.
The Strike Team is composed of researchers and engineers working to improve AI performance in complex software development scenarios, including understanding large codebases and generating complete applications. Leadership of the initiative has been assigned to Sebastian Borgeaud, previously responsible for pre training work on Gemini at Google DeepMind. The team’s work is being closely monitored and supported by senior figures including Sergey Brin and DeepMind Chief Technology Officer Koray Kavukcuoglu. Internal efforts are also being reinforced by directives encouraging greater use of AI agents for multi step coding tasks, with usage tracked through internal performance systems such as leaderboards.
This initiative is taking place within a broader competitive landscape where AI coding tools have become a major focus for leading technology companies. Anthropic has developed Claude Code as part of its early specialization in programming assistance, while OpenAI has expanded its Codex ecosystem. These developments have prompted further responses across the industry, including new design focused tools and enhancements aimed at improving developer productivity. Google’s internal strategy also includes increasing reliance on AI tools for its own engineering teams, aligning internal workflows more closely with emerging AI assisted development practices.
Industry data suggests that adoption of AI coding tools is already widespread, with surveys indicating that a large majority of developers either use or plan to use such systems. Reports show that 84 percent of developers are engaged with AI assisted coding tools, although concerns remain regarding accuracy and reliability, with nearly half of respondents expressing reservations about correctness. Google’s internal push reflects both competitive pressures and a broader industry shift toward integrating AI more deeply into software development processes, where improvements in coding capability are seen as a key step toward more advanced autonomous systems.
Follow the SPIN IDG WhatsApp Channel for updates across the Smart Pakistan Insights Network covering all of Pakistan’s technology ecosystem.
