Isomorphic Labs, a biotech company spun out from Google DeepMind, is preparing to launch human clinical trials with drugs designed using artificial intelligence powered by AlphaFold technology. This represents a watershed moment: for the first time, molecules predicted and optimized by AI will be tested in human patients, moving the technology from theoretical breakthrough to practical medicine.

The company's president, Max Jaderberg, announced the upcoming trials during a presentation in London, though he did not provide a specific timeline. The delay from initial expectations reflects the genuine complexity of transitioning from computational design to human safety testing. Last year, CEO Demis Hassabis had suggested trials would begin by the end of 2025, but the company is now taking a more cautious approach.

How Did AlphaFold Become a Drug Discovery Tool?

AlphaFold's journey from academic breakthrough to clinical application spans just a few years. In 2020, DeepMind introduced AlphaFold 2, a deep-learning system that solved one of biology's most stubborn problems: predicting how proteins fold into their three-dimensional shapes based solely on their amino acid sequences. This achievement was so significant that it earned Demis Hassabis and researcher John Jumper the 2024 Nobel Prize in Chemistry.

The impact has been staggering. Since its public release in 2021, AlphaFold has mapped roughly 200 million known protein structures, and more than 2 million researchers across 190 countries have used the tool. In 2024, the team released AlphaFold 3, which expanded beyond proteins to predict how DNA, RNA, and other molecules interact with each other, providing a more complete picture of biological systems.

What Makes AI-Designed Drugs Different From Traditional Drug Discovery?

Understanding protein structure is fundamental to drug design. Proteins are the molecular machines that carry out nearly every function in the human body. When a drug works, it does so by binding to a specific protein and altering its function. Traditional drug discovery relies on trial and error, screening thousands of compounds to find ones that bind effectively and safely. This process is expensive, time-consuming, and wasteful.

Isomorphic Labs has built on AlphaFold's foundation by creating IsoDDE, a proprietary drug design engine that goes beyond simple protein structure prediction. According to company data, IsoDDE significantly improves accuracy compared with AlphaFold 3 alone. The system can predict not only how a drug binds to its target protein but also what unintended effects might occur when the drug interacts with other proteins in the body.

This capability matters enormously for drug safety and efficacy. Jaderberg stated that AI-designed molecules may require lower doses and cause fewer side effects than traditionally discovered drugs. However, these claims will now face rigorous real-world testing. Many experimental treatments fail during human trials, and AI-designed drugs are no exception to this reality.

Steps to Understanding How AI Drug Design Works

• Protein Structure Prediction: AlphaFold analyzes the amino acid sequence of a target protein and predicts its three-dimensional shape with high accuracy, revealing where drugs can bind.
• Molecular Interaction Modeling: IsoDDE simulates how potential drug molecules would interact not just with the target protein but with hundreds of other proteins in the body, predicting side effects before synthesis.
• Optimization and Synthesis: The AI system designs multiple candidate molecules and ranks them by predicted efficacy and safety, allowing researchers to synthesize only the most promising compounds for laboratory testing.
• Preclinical Testing: Selected molecules undergo traditional laboratory and animal testing to confirm the AI predictions before advancing to human trials.

Isomorphic Labs has invested heavily to reach this clinical milestone. Last year, the company raised $600 million in funding specifically to support clinical development. It also hired a chief medical officer and expanded its clinical team to navigate the regulatory and scientific complexities of human trials.

Who Is Backing This Effort, and What Diseases Are They Targeting?

Isomorphic Labs is not working alone. The company has formed partnerships with major pharmaceutical companies including Eli Lilly and Novartis, collaborating on AI-driven drug discovery programs. These partnerships provide both credibility and resources, as established pharma companies bring decades of experience in clinical development and regulatory navigation.

The company's internal pipeline focuses on cancer and immune-related diseases, two areas where protein structure and molecular interactions are particularly critical to understanding disease mechanisms. These therapeutic areas also represent some of the largest markets in pharmaceuticals, making them attractive targets for AI-driven innovation.

The broader ambition is even more expansive. Jaderberg indicated that Isomorphic Labs aims to tackle a wide range of diseases using AI-driven design, though he acknowledged the scale of that mission. The upcoming human trials will serve as a proof of concept for whether AI can move beyond theoretical predictions and deliver practical medical outcomes that benefit patients.

Why Does This Moment Matter for the Future of Medicine?

The transition from protein prediction to human treatment represents a critical inflection point in AI's role in medicine. For years, AI has excelled at pattern recognition and prediction in controlled environments. But drug discovery requires not just accurate predictions but also the ability to navigate regulatory approval, manufacturing scale-up, and clinical efficacy in diverse human populations.

Success in these trials would validate a new paradigm: that AI can accelerate the entire drug discovery pipeline, from target identification through clinical testing. Failure would not necessarily discredit the technology but would highlight where AI predictions diverge from biological reality in living systems. Either outcome will provide crucial data for the next generation of AI-driven drug discovery tools.

Meanwhile, Google DeepMind is expanding its global footprint to support this mission. The company announced plans to open its first AI campus in Seoul, South Korea, expected to become operational within 2026. The campus will serve as a hub connecting Google engineers with South Korean startups, researchers, and industrial companies, with a focus on joint AI research in science and technology. This expansion reflects DeepMind's commitment to embedding AI development infrastructure in key markets worldwide, positioning AlphaFold and related technologies as central to global scientific progress.