Demis Hassabis, Co-Founder and CEO of Google DeepMind, offered a sobering yet electrifying assessment of the artificial intelligence trajectory, positing that the current transition will be "ten times bigger and ten times faster" than the Industrial Revolution—a 100x acceleration of change. Speaking with Bloomberg’s Emily Chang at Bloomberg House in Davos on the sidelines of the 2026 World Economic Forum, Hassabis provided sharp insights into Google’s renewed competitive position, the specific technical hurdles remaining before achieving Artificial General Intelligence (AGI), and the profound societal shifts that must accompany this era of unprecedented technological disruption.
Chang opened the discussion by asking if Google had "gotten its mojo back," referencing the competitive pressures the company faced in the early generative AI boom. Hassabis confidently acknowledged the challenge but stressed that DeepMind’s recent advances, particularly with the Gemini 3 series and the Imagen models, placed Google back at the "state of the art." He attributed this success not merely to talent, but to strategic organizational alignment and the inherent structural advantages Google possesses.
Hassabis argued that the company’s long history in foundational AI research—responsible for breakthroughs like the Transformer architecture, Deep Reinforcement Learning, and AlphaGo—combined with its unique infrastructure is an insurmountable moat. The integration of Google’s proprietary Tensor Processing Units (TPUs), expansive data centers, the robust cloud business, and ubiquitous consumer products (Search, Gmail, Chrome) creates a holistic ecosystem perfectly suited for AI deployment. Hassabis emphasized the competitive strength lies in leveraging this integrated capability: "We’re the only organization that has the full stack—from the TPUs and the hardware, the data centers, the cloud business, the frontier lab, and all of these amazing products that can, you know, kind of natural fits for AI." This structural reality allows DeepMind to operate with a speed and scale that mirrors "startup energy" while backed by institutional resources.
Regarding the timeline for true AGI, Hassabis maintained his long-standing prediction: a 50% chance of reaching the milestone by 2030. He clarified that AGI, by his definition, requires a system that exhibits all the cognitive capabilities of humans, including the ability to solve novel, complex problems. The disruption caused by achieving this level of intelligence is what warrants the "100x" comparison to historical industrial change. The true impact, he suggests, goes far beyond white-collar job displacement—it touches the very foundation of the global economy. If AGI is built correctly, it could usher in a "post-scarcity world" by solving core resource problems, such as fusion energy or the discovery of new materials.
A key focus for DeepMind now is bridging the gap between digital intelligence and physical reality—the realm of robotics. Hassabis believes the industry is on the cusp of a "breakthrough moment in physical intelligence," driven by multimodal models like Gemini, which inherently understand the physical world. However, he cautioned that this revolution still requires significant research and engineering, estimating that widespread, reliable robotics applications are still 18 to 24 months away from scalable deployment. The main obstacles are algorithmic robustness, the need for models that learn effectively with less data than their large language counterparts, and hardware limitations. Hassabis noted a renewed respect for the human body when studying robotics: "When you look into robotics very carefully, you get a newfound appreciation... for the human hand and how exquisite evolution has designed that." The dexterity and reliability of human motor skills remain the immediate benchmark.
Hassabis also addressed the geopolitical competition, specifically the perceived threat from Chinese companies following the emergence of models like DeepSeek. While acknowledging the capability of firms like ByteDance, he maintained that Chinese companies are primarily focused on "catching up" to the frontier rather than innovating beyond it. The true test for China, he asserted, is whether they can generate truly novel, frontier-pushing breakthroughs without relying on Western models.
From a purely technical standpoint, Hassabis highlighted the current flaw of advanced AI systems: "jagged intelligence." Models are excellent at specific tasks but lack consistency across the board. Achieving AGI necessitates solving for consistent reasoning, long-term planning, and, crucially, scientific creativity—the ability to formulate the right hypothesis, not just solve a known problem. Hassabis suggested that solving the scientific problem first—finding the question, which is often harder than finding the answer—is the ultimate test of true intelligence.
Given the immense power and risk inherent in AGI, the conversation turned to safety and international cooperation. Hassabis expressed a desire for a unified, global approach, suggesting the creation of an "international CERN equivalent for AI," where the world’s best minds collaborate on safety and security protocols. While he noted that informal cooperation exists among leading labs (including Anthropic), he stressed that genuine, enforceable international collaboration is necessary to manage the risks and ensure the benefits of AGI are shared globally. He concluded by affirming Google’s foundational identity as a scientific company, driven by a culture of rigorous, thoughtful, and responsible development. This scientific method, he believes, is the best defense against unforeseen risks and the only way to ensure that these powerful new tools ultimately "benefit all of humanity."
