Google DeepMind CEO: We Returned to Startup Roots After Merger

The past few years have seen Google DeepMind deliberately accelerate its operations, a strategic shift CEO Demis Hassabis credits to a powerful synthesis. By combining the vast computational resources of the former Google Brain with the pioneering research culture of DeepMind, the merged entity has returned to what he describes as a startup or entrepreneurial mindset. This cultural reset, emphasizing scrappiness, speed, and rapid deployment, is a direct response to what Hassabis calls a ferocious competitive environment. He notes that industry veterans report this period as the most intense they have ever witnessed in technology.

This acceleration is fueled by unprecedented resource alignment within Alphabet. Hassabis speaks daily with CEO Sundar Pichai, underscoring Google DeepMind’s central role in corporate strategy. That focus is backed by massive capital investment, with Alphabet’s expenditures on infrastructure nearly doubling from 2025 into 2026. The company’s custom chip development is critical here, as the primary constraint is now supply chain capacity, not available capital. This infrastructure powers what Hassabis frames as a legacy of innovation, asserting that roughly 90% of modern AI’s foundational breakthroughs originated from Google Brain, Google Research, or DeepMind. Landmark achievements like the transformer architecture and the Nobel Prize-winning AlphaFold system support this claim of historical influence.

That legacy now manifests in a sharply increased product cadence. The organization has effectively closed the gap between its frontier research and open-source contributions, with models like Gemma 4 emerging from the same infrastructure as flagship systems like Gemini 3. The latter reached about 750 million monthly active users by late 2025, with its release reportedly triggering urgent internal reviews at rival firms. This operational tempo reflects a core Hassabis thesis: that frontier research and product development are inseparable, and the entity that masters both at scale will define the industry’s future.

Concurrently, Hassabis maintains a demanding dual leadership role. After his primary workday at Google DeepMind, he begins a second workday around 10pm leading Isomorphic Labs, the group’s AI-driven pharmaceutical spinoff. He views applying AI to drug discovery as his most critical long-term ambition, requiring his direct involvement. Isomorphic, which raised $600 million in 2025 and has partnerships with Eli Lilly and Novartis worth up to $3 billion, recently launched the IsoDDE drug design tool. The company anticipates initiating human trials in oncology later this year, entering a sector where competition is intensifying. The recent acquisition of a stealth biotech startup by Anthropic signals that general-purpose AI firms now see drug discovery as a core product category, not just a research demonstration.

The strategic landscape Hassabis navigates is defined by a convergence of immense ambition and capital. While Google DeepMind leverages its unique position, the scale of investment flowing to competitors is monumental, as illustrated by SoftBank’s recent $40 billion bridge loan to OpenAI. In this context, the startup mentality Hassabis cultivates is not merely beneficial, it is an operational necessity. Institutional inertia has become a disqualifying trait in a race where the resources of incumbents and the agility of challengers are increasingly matched. The fusion of entrepreneurial speed with vast corporate resources is presented not as a temporary tactic, but as the durable advantage required to compete.