Google DeepMind, once hailed as a pioneering force in the quest for artificial general intelligence (AGI), is shifting its focus toward product development under new leadership. Koray Kavukcuoglu, the newly appointed chief of DeepMind, has signaled a clear departure from the lab’s original mission, prioritizing the release of Gemini 4 over the pursuit of AGI. Kavukcuoglu aims to launch the next-generation model “much earlier” than the end of the year, according to internal reports.
The announcement marks a significant evolution for DeepMind, which was founded by Demis Hassabis with the ambitious goal of creating machines capable of human-level intelligence. However, Kavukcuoglu has dismissed the AGI conversation as “not the right conversation,” instead emphasizing the importance of building trustworthy AI agents. The model is already undergoing post-training and is being tested within Google’s internal coding platform, Antigravity.
This pivot comes amid a wave of talent exodus from DeepMind, with several top researchers moving to competing firms like OpenAI and Anthropic. The lab’s shift from pure research to product development reflects broader industry trends, where companies are increasingly focused on delivering commercial AI solutions rather than pursuing long-term theoretical goals. As the AI landscape becomes more competitive, DeepMind’s new direction may be a strategic move to maintain relevance and market position.
With the quiet disappearance of Gemini 3.5 Pro and a leadership change, DeepMind’s identity is evolving. While the lab’s original vision of AGI remains a tantalizing possibility, the immediate priority now lies in bringing powerful, reliable AI tools to market—starting with Gemini 4.