No Drugs, Yet Google’s AI Pharma Startup Seeks $5 Billion Funding

A Drug-Free Google AI Pharma Company Seeks $5 Billion Funding

On November 8, Bloomberg reported that the company is negotiating a new financing round of about $5 billion, which could value it at $50 billion.

That figure stands out dramatically within the entire AI pharma industry. The global AI drug discovery market is projected to be about $2.3 billion in 2025, and this company still has no drug that has entered human clinical trials.

Just a few days ago, Isomorphic Labs President Max Jaderberg publicly predicted that all future drugs would be designed by AI.Now, this bold claim linked with the massive financing news has drawn even more attention to the commercial prospects of AI pharma.

Nobel laureate joins, cross‑disciplinary team supports drug development

Why can a company with no drugs yet in clinical trials command such a high valuation? Beyond the market potential of AI‑driven drug discovery itself, the team of drug‑development experts behind Isomorphic Labs may also be a key reason investors are bullish.

Chief Scientific Officer Miles Congreve has over 25 years of drug‑development experience, has helped design multiple drugs that entered clinical evaluation, and is a co‑inventor of the approved breast‑cancer drug Kisqali.

Chief Research Officer Sarah Skerratt oversees research strategy and the drug discovery pipeline, Chief Medical Officer Ben Wolf drives candidate drugs into clinical trials, and Chief Technology Officer Sergei Yakneen brings over 20 years of engineering and machine‑learning experience.

The company’s Scientific Advisory Board also includes CRISPR gene‑editing pioneer and 2020 Nobel Chemistry laureate Jennifer Doudna, as well as 2021 Nobel Chemistry laureate David MacMillan and other leading scientists.From early molecular design to clinical development, this team spans multiple critical stages of drug R&D.

Image | The company’s advisory board (source: Isomorphic Labs)

In addition to experience in traditional pharmaceuticals, Isomorphic Labs also has a strong background in artificial intelligence research. The company’s founder and CEO Demis Hassabis also serves as CEO of Google DeepMind, and earned the 2024 Nobel Prize in Chemistry for his AlphaFold-related research.

The company’s day-to-day operations are overseen by President Max Jaderberg. He conducted computer vision and machine learning research at Oxford University, then led the Open-Ended Learning team at DeepMind, and was involved in projects such as AlphaStar.

Image | The company’s executive team (Source: Isomorphic Labs)

This also creates an important characteristic of Isomorphic Labs.On one hand, it inherits DeepMind’s experience in AI model development; on the other hand, it has recruited professionals familiar with drug discovery and clinical development.However, whether the combination of technology and talent can ultimately translate into tangible results still depends on navigating the most challenging stages of drug development.

AI drug discovery: can it overcome clinical hurdles?

AI can accelerate early-stage molecular screening, but the real challenge remains the long and costly clinical trials. In traditional drug development, about 90% of candidate drugs that enter human testing ultimately fail, due to toxicity and side effects or because they do not achieve the expected efficacy.

Jaderberg has his own take on this. He believes that traditional pharmaceutical methods rely too heavily on trial and error, requiring researchers to sift through millions of molecules to find candidates that can bind to a target. This approach limits the complexity of drug design, and over the past few decades the pharma industry has largely been developing targets that are relatively easy to achieve.

According to Jaderberg, AI is expected to transform this process, allowing drug discovery to be built on more comprehensive computational predictions. In a recent talk, he described two complementary types of AI models: the first type is inference models that help researchers generate scientific hypotheses; the second type is simulation models that can approximate the laws of the physical world on a computer, predicting molecular behavior in advance.

Combining the two model types enables researchers to evaluate more design options before conducting actual experiments, reducing futile attempts and continuously optimizing the structures of candidate molecules. However, the extent to which this approach can improve the success rate of drug development still requires validation with real experimental data.

According to the established plan, Isomorphic Labs expects to launch its first human clinical trial by the end of 2026.If implemented smoothly, the AI‑designed drugs developed in‑house will enter the clinical validation stage. However, early‑stage trials mainly focus on safety and tolerability, and there is still a long way to go before the drug’s true efficacy can be proven.

In‑house development and partnerships run in parallel, with IsoDDE as the technical core

While awaiting clinical validation, Isomorphic Labs has already begun advancing its business through both independent R&D and collaborations with pharmaceutical companies.Jaderberg previously made it clear that the company does not sell APIs nor Tokens; the ultimate purpose of its AI model IsoDDE is to develop real drugs.

Currently, the company has two main business tracks. One is independent R&D, where the team determines the drug development direction, uses AI to design candidate drugs, and invests resources to advance development, focusing on cancer, immune system diseases, and inflammation.

The other track involves collaborations with large pharma companies such as Eli Lilly, Novartis, and Johnson & Johnson, where partners provide disease biology knowledge and R&D targets, and Isomorphic Labs is responsible for translating them into candidate molecules, earning revenue through upfront payments, milestone payments, and potential future sales royalties.

Both tracks rely on the company’s core technology platform IsoDDE. In February this year, Isomorphic Labs publicly disclosed part of the performance of this AI drug design engine, showcasing its advances in protein structure and molecular interaction prediction.

As one of the technical foundations, AlphaFold 3 can already predict the interaction structures between proteins and other molecules, but its prediction accuracy may still drop when faced with new targets or molecules that are insufficiently represented in the training data. IsoDDE aims to improve the model’s generalization ability on these unknown systems.

According to the technical report released by Isomorphic Labs, on the Runs N’Poses benchmark specifically designed to test generalization, when dealing with complex systems that differ significantly from the training data, IsoDDE’s structural prediction accuracy exceeds that of AlphaFold 3 by more than twofold.

Figure | In the Runs N’Poses benchmark, protein‑ligand structure prediction performance in the hardest generalization category (0,20] (source: Isomorphic Labs)

For predictions of antibody‑antigen binding interfaces, IsoDDE also outperforms AlphaFold 3 and the comparable open‑source model Boltz‑2 in the high‑accuracy range.

In addition, IsoDDE can predict the binding affinity between a molecule and its target protein. Isomorphic Labs claims that the system can make predictions at a speed far faster than traditional physics‑based simulations without relying on experimentally resolved crystal structures, with accuracy approaching that of high‑precision physics simulation methods.

Image | IsoDDE drug design engine (source: Isomorphic Labs)

Computational biologist Mohammed AlQuraishi at Columbia University in the United States said after reading the technical report that IsoDDE excels at predicting molecular and protein interactions that differ significantly from the training data, suggesting the team may have achieved novel technical breakthroughs.

However, he also noted that the report does not disclose enough technical details, so outsiders cannot independently reproduce the results at this time. In response, Jaderberg said the company has no plans to release the core technology behind it, but hopes the report will spur other teams to continue exploring AI-driven drug discovery.

For a company that has yet to have any drug enter human trials, a potential valuation of $50 billion reflects a market bet on a commercial premise:whether AI can improve the accuracy of drug design, reduce futile attempts in R&D, and ultimately increase the success rate of new drug development.

Jaderberg once likened the company’s development stage to a baseball game, saying the team is still in the first inning. The massive financing can provide ample funds for subsequent R&D, but whether this bet will ultimately pay off will depend on the clinical trial results that will be gradually released over the next few years.

Reference materials:

1.https://www.bloomberg.com/news/articles/2026-11-04/isomorphic-labs-in-talks-for-about-5-billion-in-funding

2. https://www.isomorphiclabs.com/our-team

3.https://www.isomorphiclabs.com/articles/meet-our-team-agnieszka-michael