Google Pushes Gemini AI Overhaul as Sergey Brin Targets Model Supremacy

Google co-founder Sergey Brin has stepped up efforts to accelerate development of the company’s Gemini AI models as Alphabet battles OpenAI and Anthropic for leadership in artificial intelligence.

Brin has reportedly urged key Google DeepMind employees to move faster and focus resources on reaching the AI frontier. During an April town hall, he encouraged hundreds of staff members to intensify development as Anthropic prepared advances to its Claude models.

Google briefly gained an edge over competitors with Gemini 3 last November, but subsequent releases from OpenAI and Anthropic increased pressure on the company. By August, Google had reportedly delayed its next flagship Gemini model by two months after internal testing showed weaknesses in areas including coding.

The competitive challenges have coincided with major changes at Google DeepMind. On August 5, Demis Hassabis moved from chief executive to chair, while deputy Koray Kavukcuoglu took control of the AI division. Kavukcuoglu, who also serves as Google’s chief AI architect, has increasingly shaped Gemini development and its integration across Google products.

The restructuring also shifted several teams from DeepMind into Google’s broader corporate organization, raising concerns among employees about the AI lab’s declining autonomy. Some staff reportedly worry that commercial priorities could increasingly replace the freedom to pursue independent research.

Brin, despite holding no formal executive position, has also pushed resources toward areas such as recursive self-improvement, where AI systems could potentially improve their capabilities with reduced human involvement.

Google’s difficulties have included limited access to Tensor Processing Units, internal disagreements over resource allocation and slower product releases compared with smaller AI rivals. CEO Sundar Pichai has said Alphabet plans to expand infrastructure to address computing constraints.

The leadership overhaul also strengthens Kavukcuoglu’s position in Mountain View and could improve coordination between DeepMind and Google Cloud, a major source of AI revenue.

With Gemini central to Google’s search, cloud and consumer-product strategy, the changes signal that Alphabet remains firmly committed to the AI model race. Rather than retreating, Google appears to be reorganizing DeepMind around a clearer objective: making Gemini competitive with the world’s most advanced AI models.