Google Restructures AI Teams Amid Pressure to Advance Gemini Models

​Google is intensifying efforts to regain its competitive edge in artificial intelligence. Co-founder Sergey Brin has urged employees to prioritize AI development as Google seeks to close the gap with rivals Anthropic and OpenAI. Recent reports indicate Brin has encouraged key engineers to accelerate progress on the Gemini AI models, which have lagged behind competitors in recent months.

Key Highlights

• Sergey Brin urges Google staff to accelerate Gemini AI model development.
• DeepMind CEO Demis Hassabis steps down from daily management responsibilities.
• Gemini app surpasses 1 billion users but lags in enterprise adoption.
• Four key Google scientists leave to start their own AI company.
• Google restructures DeepMind, moving non-technical teams into corporate reporting.

Leadership Push for Gemini Advancement

Brin addressed hundreds of Google DeepMind employees at a town hall in April 2024, following Anthropic's release of the Claude Mythos model. At the time, Claude Mythos was considered the most powerful AI model available. Brin called on staff to speed up work on Gemini, emphasizing the urgency to catch up with competitors. Although Brin does not hold an executive position, he continues to influence Google's AI model training efforts.

Gemini has achieved rapid global adoption, with the Gemini app surpassing 1 billion users. However, it trails behind OpenAI and Anthropic in attracting enterprise customers. Enterprise users are crucial, generating billions in revenue for leading AI firms. Google delayed the release of its next major model, Gemini 3.5 Pro, after it failed to meet internal coding goals. This delay has further widened the gap with competitors.

Organizational Changes at DeepMind

Google is undergoing significant internal changes as it seeks to strengthen its AI initiatives. Four key scientists recently left the company to launch their own AI startup. In addition, DeepMind, Google's advanced AI lab, is experiencing a leadership transition. Demis Hassabis, DeepMind's CEO, has stepped down from daily management duties. Koray Kavukcuoglu, a senior vice president at DeepMind, will assume Hassabis's responsibilities.

Following the leadership change, Hassabis and Kavukcuoglu addressed staff at an all-hands meeting. They assured employees that daily operations would remain unchanged and that Kavukcuoglu would continue overseeing Gemini. However, several non-technical teams are moving from DeepMind to Google's broader corporate structure. This marks a further reduction in DeepMind's autonomy since Google acquired the London-based lab in 2014.

The reorganization has raised concerns among some employees about DeepMind's independence. The shift signals Google's intent to focus more directly on regaining leadership in AI model development. Influence within the Gemini project has also shifted, with some London-based leaders feeling reduced control after Kavukcuoglu relocated to Google's California headquarters and Hassabis stepped back from the project.

Competitive Landscape and Future Outlook

Google had narrowed the gap with competitors in November 2023 with the release of Gemini 3. However, it now trails newer models like Anthropic’s Fable 5 and OpenAI’s GPT-5.6 Sol. Other companies, including Meta and SpaceXAI, have also advanced their AI models. Chinese firms such as Moonshot, with its Kimi K3 model, are making progress as well.

Google’s recent organizational changes and leadership initiatives reflect the company’s determination to reclaim a leading position in the AI sector. The company continues to adapt its structure and strategy to meet the evolving demands of AI development and competition.