Lead
Sir Demis Hassabis, CEO of Google DeepMind and 2024 Nobel Prize laureate in Chemistry, told Fortune on Feb. 11, 2026 that AI could usher in a “new golden era of discovery” within 10 to 15 years. He said the transition will be turbulent, requiring major companies such as Google to disrupt their own businesses to capture long-term gains. Hassabis pointed to breakthroughs like AlphaFold and recent model releases as evidence that AI can accelerate research in medicine, energy and materials. The short-term result, he warned, is a decade-long shakeout across the $3.9 trillion firm and the broader industry.
Key Takeaways
- Hassabis forecasts a 10–15 year “renaissance” driven by AI, predicting major changes in medicine, materials and exploration.
- He frames the coming decade as a disruptive sprint for Google, a $3.9 trillion company that may need to cannibalize search to scale generative AI.
- DeepMind merged Google Brain and DeepMind in 2023 to pool compute and talent; Hassabis likens the combined unit to a “nuclear power plant” feeding products across Alphabet.
- AlphaFold — DeepMind’s protein-structure system — has predicted roughly 200 million protein structures and is used by over 3 million researchers, the accomplishment behind Hassabis’s 2024 Nobel Prize in Chemistry.
- Isomorphic Labs, a Google spinoff led by Hassabis, is pursuing in silico drug discovery and reports preclinical cancer programs with hopes to enter clinical trials by year-end.
- Alphabet’s stock rallied about 65% following recent model launches (e.g., Gemini 3) and viral product demonstrations like Nano Banana, reflecting investor optimism.
- Hassabis warns that failing to self-disrupt invites competitors to reshape core businesses; he said, “If we don’t disrupt ourselves, someone else will.”
Background
The rise of generative AI since 2022 has forced legacy internet companies to reassess priorities that once centered on search and advertising. OpenAI’s ChatGPT and other large language models demonstrated new consumer and enterprise use cases, accelerating product races and prompting reorganizations across Silicon Valley. In 2023 Google combined its premier research arms—Google Brain and DeepMind—into a single organization under Hassabis to concentrate compute, talent and research direction.
DeepMind’s AlphaFold, unveiled earlier in the decade, marked a milestone by solving the long-standing protein-folding problem and producing an open database of predicted protein structures. That output has been widely adopted by the life-science community, informing basic biology and drug research. At the same time, investors have reacted strongly to new product releases and demos, valuing firms that appear to commercialize frontier models quickly.
Main Event
Speaking on Fortune’s podcast on Feb. 11, 2026, Hassabis painted a two-stage picture: a volatile near term followed by an extended period of accelerated discovery. He argued that building general-purpose scientific assistants requires risking established revenue streams, and that Google’s internal reorganization was designed to enable exactly that. Hassabis described the merged research unit as supplying intelligence to Search, YouTube and other products while pursuing high-end scientific applications.
He highlighted AlphaFold’s impact—about 200 million predicted protein structures and adoption by more than 3 million researchers—as a concrete example of AI turning foundational science into practical tools. That success, he said, is the template for scaling AI across other scientific domains, from materials design to fusion research. Hassabis also referenced Isomorphic Labs’ pipeline: moving discovery from wet labs to simulation to increase throughput, with preclinical cancer programs progressing toward clinical trials by year-end.
Hassabis framed the coming decade as a necessary period of upheaval. He described long work hours and intense corporate changes—reorgs, compute investments and product pushes—as the price of reshaping an industry. The immediate consequence has been heightened competition, sharper investor reactions and a wave of productization that tests existing business models, especially search.
Analysis & Implications
If Hassabis’s timeline holds, the next 10–15 years could see fundamental shifts in how science is conducted. AI systems that reliably propose experiments, design molecules and interpret complex datasets would shorten iteration cycles and reduce costs in drug development, materials research and energy. Economically, this could translate into new high-value sectors and displace legacy workflows, altering labor demand in R&D and accelerating capital deployment into compute and specialized instrumentation.
For Google and Alphabet, the dilemma is strategic: prioritize steady ad-backed search revenues or pivot aggressively toward research-driven platforms that may not produce immediate profit. The 2023 consolidation of Brain and DeepMind reflects a deliberate bet on long-term platform and product integration; it also concentrates risk if frontier models fail to deliver practical advantages quickly. Investor responses—such as the roughly 65% stock lift tied to recent model releases—signal market willingness to reward perceived technological leadership, but also raise expectations.
Scientifically, AlphaFold’s footprint shows how open, high-quality model outputs can bootstrap entire research communities. If similar gains occur in computational materials, energy systems or automated lab design, the pace of discovery could accelerate nonlinearly. However, bottlenecks will remain: data quality, experimental validation, regulation for clinical translation, and the capital intensity of large-scale compute.
Comparison & Data
| Metric | Reported Value |
|---|---|
| Alphabet market footprint referenced | $3.9 trillion |
| AlphaFold predicted structures | ~200 million proteins |
| Researchers using AlphaFold | Over 3 million |
| Alphabet share price change after recent releases | ~+65% (by year-end) |
| Hassabis projected timeline | 10–15 years |
The table aggregates figures cited by Hassabis and reported in the interview: company scale ($3.9 trillion), AlphaFold reach (200 million structures; 3 million users), and market reaction (~65% share gain). These data points frame both the scientific impact and commercial stakes of DeepMind’s work. They should be read as illustrative of scale rather than precise causal proof of near-term outcomes.
Reactions & Quotes
Google and DeepMind stakeholders framed the reorg and product releases as strategic moves to integrate research and engineering at scale. Observers in industry note the tension between short-term monetization and long-term research goals, and regulators are watching how AI’s role in health and critical infrastructure evolves.
If we don’t disrupt ourselves, someone else will.
Demis Hassabis, CEO, Google DeepMind (interview)
The remark summed up Hassabis’s rationale for the 2023 consolidation of Brain and DeepMind and for taking risks that could affect Google’s search franchise. He used the line to underscore the urgency of self-directed disruption as a defensive and offensive strategy.
We set out with the mission of solving intelligence and then using it to solve everything else.
Demis Hassabis (interview)
This succinct formulation connects DeepMind’s long-term research mission to its current push into applied science, from AlphaFold to Isomorphic Labs. It also frames the broader corporate aim: to translate foundational AI capabilities into tools for real-world problems.
AlphaFold’s database has become a road map for biology used by millions of researchers.
Industry analysis / public reporting
Independent researchers and partner institutions have cited AlphaFold as a major accelerator for structural biology. That endorsement is one reason the model’s output is often referenced as the closest concrete example of AI materially speeding discovery.
Unconfirmed
- The claim that in silico methods will routinely be “1,000 times” more efficient than traditional wet-lab workflows lacks independent, peer-reviewed quantification at scale.
- Precise timing of Isomorphic Labs moving specific cancer programs into clinical trials by year-end is an aspiration reported by the company but subject to regulatory and scientific milestones.
- Long-range visions—such as AI enabling interstellar travel—are speculative and depend on breakthroughs in energy, propulsion and materials beyond current validated demonstrations.
Bottom Line
Hassabis offers a high-confidence vision: AI can become a force-multiplier for science and industry within a 10–15 year window, but reaching that point requires a decade of organizational and technological turbulence. For Google and other major players, the central choice is whether to accept near-term disruption in exchange for leadership in a redefined research economy. Investors have already signaled enthusiasm, but concrete societal benefits will hinge on validated scientific outcomes, regulatory pathways and equitable access to the technologies.
Readers should watch three things closely over the coming years: the pace at which computational discoveries translate into validated clinical or materials outcomes; how companies balance core businesses with experimental bets; and how regulators adapt to faster, AI-driven R&D cycles. If Hassabis’s timeline proves broadly accurate, the coming decade will be decisive in shaping both the winners and the ethical frameworks around powerful scientific AI.