DeepMind has a new leader who faces a significant challenge in enhancing the company's AI performance relative to key competitors, OpenAI and Anthropic. Recently, both of these firms have launched advanced systems that have garnered attention and sparked debate due to their impressive capabilities. Meanwhile, Google, which acquired DeepMind in 2014, has not introduced a groundbreaking model since early 2026. The newly appointed leader, Koray Kavukcuoglu, who previously held the positions of CTO for the AI division and Google's chief AI architect, is tasked with accelerating progress in developing the industry's top models. He takes over as Senior Vice President from DeepMind co-founder and former CEO Demis Hassabis, who will now serve as chair of the organization. According to Ben Wood, chief analyst at CCS Insight, the expectation is clear: Kavukcuoglu will aim to narrow the gap with Anthropic and OpenAI, shifting Google's focus from academic endeavors to enhancing performance at the frontline and crafting better tools for developers.
In terms of competition, Google has found itself trailing OpenAI after the latter released ChatGPT to the public in 2022. Although Google made a notable comeback with Gemini 3 in November 2025—an update that analysts praised for pushing the limits of frontier models—2026 has seen OpenAI and Anthropic advance further with new offerings, including Anthropic’s Mythos model and OpenAI's GPT-5.6.
Since launching Gemini 3.1 Pro in February of this year, Google has not introduced models that significantly challenge the frontier. To enhance its coding capabilities, pivotal in the pursuit of artificial general intelligence (AGI), the company has established an internal team known as "Code Strike." Malik Ahmed Khan, a senior equity analyst at Morningstar, pointed out that Google's concentration on profit-generating sectors like search and multi-modal applications caused it to miss out on the vital coding use case, where its rivals currently enjoy a distinct advantage. A spokesperson for Google clarified that the company is fully committed to its broader mission of solving "intelligence" as a priority above commercial interests, emphasizing that advancing AGI will entail progress in various areas, including agents, coding, robotics, and world models.
As Kavukcuoglu assumes his new role, he emerges from DeepMind's original team, having joined in 2012 before the company’s acquisition by Google. His goal will be to revitalize the challenge posed to Anthropic and OpenAI, especially in light of recent high-profile exits from the company. Reporting directly to Google CEO Sundar Pichai, Kavukcuoglu will oversee the development of the Gemini model, lead Frontier AI research, and manage the Gemini app and developer teams. Analysts view his promotion as a signal that Google aims to direct more energy towards enhancing large language models and that this strategic shift positions the company more favorably in the race for advanced AI development.
Hassabis’ leadership at DeepMind included a broad emphasis on areas like multi-modality and health-focused projects, as evidenced by initiatives like Isomorphic Labs. Over the past year, Kavukcuoglu has taken on increased responsibilities in guiding model development and showcasing key Gemini releases, indicating a potential pivot in Google's strategy towards more rapid execution on product development rather than deep academic inquiry.
After a disappointing report in July regarding delayed plans for the Gemini 3.5 Pro model—causing Alphabet's stock to drop—analysts suggest that the primary objective for Kavukcuoglu should be to successfully launch this model and establish a reliable release schedule moving forward. Furthermore, it is crucial for him to re-establish expertise in coding and pretraining amidst departures from the company.
Previously headquartered in London, DeepMind retained its presence there under Hassabis' leadership. As Kavukcuoglu transitioned from London to Google's main campus in Mountain View, California, some experts noted that Google seems to be consolidating its AI leadership outside the U.K. However, a company spokesperson emphasized their commitment to maintaining operations in the UK capital.
