NextFin News - Demis Hassabis, the Nobel Prize-winning CEO of Google DeepMind and Isomorphic Labs, has intensified his efforts to transform the global pharmaceutical landscape by applying advanced artificial intelligence to the fundamental challenges of biology. As of late January 2026, Isomorphic Labs, an Alphabet-owned subsidiary, has transitioned from theoretical protein folding to active drug design, targeting high-stakes therapeutic areas including oncology and immunology. According to Bloomberg, Hassabis aims to reduce the traditional drug discovery phase—which typically spans three to six years—down to a matter of months or even weeks. This acceleration is being facilitated through the deployment of AlphaFold 3 and subsequent proprietary models that simulate how complex molecules, RNA, and proteins interact within the human body.
The operational scale of this ambition is underscored by high-profile commercial collaborations. Isomorphic Labs has recently expanded its partnership with Novartis, increasing its joint drug-target portfolio from three to six specific projects. Simultaneously, a multi-year agreement with Eli Lilly, valued at up to $1.7 billion in potential milestones, is leveraging Isomorphic’s platform to identify small-molecule therapies for undisclosed targets. While no AI-designed drug from the Isomorphic pipeline has yet completed the full gauntlet of Phase III clinical trials, the company has recently bolstered its clinical expertise by appointing Ben Wolf as Chief Medical Officer to oversee the transition from digital discovery to human testing. This move signals a critical pivot for the London-based startup as it seeks to prove that its computational breakthroughs can survive the biological unpredictability of the clinic.
The drive behind Isomorphic Labs represents a direct assault on "Eroom’s Law"—the pharmaceutical industry’s observation that drug discovery is becoming slower and more expensive over time despite technological gains. Historically, the cost of developing a new drug has doubled roughly every nine years, reaching an estimated $2.6 billion per approved molecule. Hassabis is betting that AI can invert this trend by shifting the "trial and error" process from the wet lab to the silicon chip. By utilizing "digital twins" of molecular structures, researchers can discard millions of non-viable compounds before a single pipette is touched. Data from Nature Biotechnology suggests that AI-discovered drugs are already showing an 80-90% success rate in Phase I trials, significantly higher than the industry average of 40-65%, indicating that AI is not just faster, but more accurate in selecting candidates with favorable toxicity profiles.
However, the path forward is fraught with structural and biological hurdles. While AI can predict how a drug binds to a protein, it still struggles to model the systemic complexity of the human body—the so-called ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) properties. To address this, Hassabis has noted that Isomorphic is working on "several AlphaFold-level breakthroughs" in parallel, moving beyond simple structure prediction into dynamic chemical simulation. This multi-modal approach is essential because the regulatory environment, overseen by the U.S. Food and Drug Administration (FDA), remains rooted in empirical human data. Even if discovery is shortened to months, the clinical and regulatory phases still require years of observation. Under the current administration, U.S. President Trump has emphasized streamlining federal regulations, which could potentially accelerate the approval pathways for AI-driven medical breakthroughs if they demonstrate superior safety metrics.
Looking ahead, the success of Isomorphic Labs will likely redefine the "Full-Stack" biotech model. Rather than merely licensing software, Hassabis is positioning Isomorphic as a co-developer that shares in the long-term upside of successful therapies. If the company can successfully move its internal oncology candidates into Phase II trials by 2027, it will validate the thesis that AI is the primary engine for the next generation of precision medicine. The broader impact will be a shift from reactive medicine to a predictive model where treatments are designed for the specific molecular signatures of a disease. As Hassabis continues to bridge the gap between AGI research at DeepMind and applied biology at Isomorphic, the pharmaceutical industry faces a definitive choice: integrate these computational frameworks or risk obsolescence in an era where the speed of innovation is no longer limited by the pace of the laboratory, but by the scale of the compute.
Explore more exclusive insights at nextfin.ai.