Episode 62: Isomorphic Labs

In this episode of Petri Dish Perspectives, we explore the story of Isomorphic Labs, Alphabet's ambitious drug discovery company born out of DeepMind's groundbreaking work on AlphaFold. From solving one of biology's greatest mysteries, protein folding, to building an AI-native pharmaceutical company, Isomorphic Labs is attempting to fundamentally change how medicines are discovered.

We'll dive into the origins of AlphaFold, the scientific vision of Nobel Prize-winning CEO Demis Hassabis, how AI is reshaping medicinal chemistry, and why pharmaceutical giants like Novartis and Eli Lilly are investing billions in AI-powered drug discovery partnerships. We'll also discuss the promise, challenges, and future of foundation models in biology, and whether AI can truly accelerate the journey from molecule to medicine.

Because the next blockbuster drug may not be discovered by chance, it may be designed by an algorithm.

🎧 Listen now, stay curious, and don’t forget to subscribe for new episodes every Thursday!

https://linktr.ee/maneadkhin

#PetriDishPerspectives #IsomorphicLabs #DeepMind #AlphaFold #ArtificialIntelligence #AI #DrugDiscovery #Biotech #Healthcare #PrecisionMedicine #Pharmaceuticals #MachineLearning #ProteinFolding #DemisHassabis #NobelPrize #MedicinalChemistry #LifeSciences #Biology #Google #Alphabet #EliLilly #Novartis #FutureOfMedicine #SciencePodcast #BiotechPodcast

Transcript

Hello and welcome to Petri Dish Perspectives, the podcast where we geek out about science and the companies shaping the future of healthcare. I’m your host, Manead, and I’m a PhD scientist by training, storyteller by choice. With every new episode released on Thursday, my goal is to deliver digestible pieces of information on healthcare companies under 30 mins. 

Artificial intelligence has promised to revolutionize healthcare for decades. Every few years, another startup claims it can discover drugs faster, cheaper, and more intelligently than traditional pharmaceutical companies. Most fail. Drug discovery remains one of the most expensive and risky industries on Earth, with more than 90% of experimental medicines never reaching patients.

But in 2021, one company emerged with a very different ambition. It wasn't founded by venture capitalists looking for the next unicorn, nor by pharmaceutical veterans trying to modernize old workflows. Instead, it was born inside one of the world's most advanced AI research organizations: DeepMind.

The company is Isomorphic Labs, a subsidiary of Alphabet, and its mission is remarkably bold—to solve disease using artificial intelligence. Not simply to help researchers identify better drug targets or analyze clinical trial data, but to fundamentally rethink how medicines are invented from the atomic level upward. Just in May 2025, Isomorphic had a massive $2.1 billion Series B financing round. 

Quick disclaimer, I give full credit to the original articles cited in the references in the transcript!

Grab a coffee or tea, settle in, and let’s jump in!

The Scientific Origins: Why Protein Folding Was Biology's Greatest Puzzle

To understand Isomorphic Labs, we first have to understand one of biology's oldest unsolved problems.

Every cell in your body contains DNA, but DNA doesn't actually perform most biological functions. Instead, DNA serves as an instruction manual for building proteins. These proteins act as enzymes, receptors, antibodies, hormones, and structural components that keep life functioning.

The problem is that proteins begin as simple chains of amino acids. After they're synthesized, they rapidly fold into incredibly complex three-dimensional shapes. That shape determines everything—whether the protein binds another molecule, activates a signaling pathway, or causes disease.

For decades, scientists could determine protein structures experimentally using X-ray crystallography, nuclear magnetic resonance spectroscopy, or cryo-electron microscopy. These techniques were extraordinarily accurate but painfully slow, expensive, and technically difficult. Some proteins took years to solve, while others resisted characterization entirely.

Scientists often referred to protein folding as biology's "grand challenge." If researchers could reliably predict a protein's final structure directly from its amino acid sequence, drug discovery could accelerate dramatically.

For over fifty years, this remained largely unsolved.

DeepMind and AlphaFold: The Breakthrough That Changed Biology

The turning point came from an unexpected place.

DeepMind had already become famous after building AlphaGo, the artificial intelligence system that defeated world champion Go player Lee Sedol in 2016. Many assumed the company would continue focusing exclusively on games and general artificial intelligence.

Instead, DeepMind quietly redirected some of its brightest researchers toward biology.

Led by Demis Hassabis and an interdisciplinary team of computational scientists, the company developed AlphaFold—an AI model capable of predicting protein structures with unprecedented accuracy.

The breakthrough came during the Critical Assessment of Structure Prediction competition, commonly known as CASP. In 2020, AlphaFold2 achieved a level of accuracy that many structural biologists had previously believed impossible.

Rather than improving predictions by small margins, AlphaFold effectively solved one of biology's grand challenges.

Within two years, DeepMind and EMBL-EBI released predicted structures for over 200 million proteins, covering nearly every known protein sequence on Earth. Suddenly, researchers studying cancer, infectious diseases, neuroscience, rare genetic disorders, and plant biology gained access to structural information that previously would have required decades of laboratory work.

The impact was immediate. Thousands of scientific papers began citing AlphaFold. Universities integrated it into research workflows. Pharmaceutical companies adopted it across drug discovery programs. In 2024, Demis Hassabis and John Jumper were awarded the Nobel Prize in Chemistry alongside David Baker for breakthroughs in computational protein science.

Yet despite this extraordinary achievement, DeepMind viewed AlphaFold not as the destination, but merely the first step.

Founding Isomorphic Labs: Building an AI-Native Pharmaceutical Company

Google DeepMind developed the AlphaFold AI model, which revolutionized structural biology by accurately predicting how proteins fold. Isomorphic Labs is an Alphabet-owned commercial spin-off created from DeepMind to utilize this breakthrough, specifically applying and expanding AlphaFold's technology for commercial drug discovery.

In late 2021, Alphabet officially launched Isomorphic Labs.

The name "isomorphic" comes from mathematics, referring to different systems that share the same underlying structure. For the company, it symbolizes the belief that biology itself follows computational principles that artificial intelligence can eventually understand.

Demis Hassabis became CEO while continuing to lead DeepMind.

Unlike traditional biotechnology startups, Isomorphic Labs did not begin with a single drug candidate or disease area. Instead, the company set out to build what it calls a "general-purpose drug discovery engine."

Rather than asking scientists to manually identify promising molecules, Isomorphic wants AI to design medicines from first principles.

This represents a fundamental shift.

Historically, drug discovery has been largely empirical. Researchers screen millions of compounds, optimize chemistry through repeated experiments, and gradually improve molecules over years.

Isomorphic believes artificial intelligence can dramatically compress this process by modeling molecular interactions computationally before compounds ever enter the laboratory.

Instead of searching blindly through chemical space, AI could intelligently navigate trillions of possibilities.

The Technology Platform: Beyond AlphaFold

Many people mistakenly believe Isomorphic Labs simply commercializes AlphaFold.

That isn't true.

AlphaFold predicts protein structures. Drug discovery is far more complicated.

Scientists must understand how proteins move, how they interact with small molecules, how mutations alter function, how water molecules influence binding, and how candidate drugs behave inside living systems.

Isomorphic is developing AI models capable of simulating entire molecular systems rather than static protein structures.

These models incorporate structural biology, medicinal chemistry, molecular dynamics, protein-ligand interactions, quantum chemistry, and increasingly, generative AI.

Instead of asking, "What does this protein look like?" the platform asks much deeper questions.

Which pocket is druggable?

How can we improve binding affinity?

Will this molecule avoid off-target toxicity?

Can we optimize oral bioavailability before synthesis?

Can AI generate an entirely new molecular scaffold that no chemist has previously imagined?

These questions move far beyond prediction into genuine invention.

The Pharmaceutical Partnerships: Novartis and Eli Lilly

In 2024, Isomorphic Labs announced two landmark partnerships that immediately captured the industry's attention.

The first was with Novartis.

The second was with Eli Lilly.

Combined, the agreements carried potential milestone payments approaching three billion dollars.

Both collaborations focus on applying Isomorphic's AI platform to multiple therapeutic programs across oncology and immunology.

Importantly, these deals are structured as true research partnerships rather than software licensing agreements.

The pharmaceutical companies contribute decades of disease expertise, medicinal chemistry, clinical development, and regulatory capabilities.

Isomorphic contributes artificial intelligence capable of generating and optimizing novel drug candidates.

These partnerships represent a broader shift occurring across the industry.

Pharmaceutical companies increasingly recognize that future competitive advantage may depend less on traditional laboratory throughput and more on computational design.

Rather than replacing scientists, AI becomes another member of the discovery team.

As of May 2026, Isomorphic Labs closed a massive $2.1 billion Series B funding round led by Thrive Capital. This is a critical update, as it moves the company from an "Alphabet subsidiary" to a powerhouse with a nearly $10 billion valuation and significant independent capital. 

People Who Made Their Mark: Demis Hassabis

No discussion of Isomorphic Labs is complete without Demis Hassabis.

Few individuals have influenced artificial intelligence and biology simultaneously to the degree that he has.

Hassabis began as a child chess prodigy before becoming one of Britain's youngest professional video game developers. He later earned a PhD in neuroscience, studying memory, imagination, and intelligence at University College London.

These seemingly unrelated experiences—gaming, neuroscience, psychology, and computer science—eventually converged into DeepMind.

His vision has always extended beyond building smarter algorithms.

He believes artificial intelligence should become a scientific instrument capable of accelerating human discovery itself.

That philosophy ultimately produced AlphaFold, Isomorphic Labs, and one of the most ambitious attempts to reinvent pharmaceutical R&D.

In 2024, Hassabis shared the Nobel Prize in Chemistry, cementing his place as one of the defining scientific leaders of this generation.

Challenges and Criticisms

Despite enormous enthusiasm, Isomorphic Labs faces significant challenges.

The first is validation.

Predicting molecules computationally is fundamentally different from proving they work in humans. Biology remains extraordinarily complex, and even perfectly designed molecules can fail because of unexpected toxicity, poor pharmacokinetics, or unforeseen clinical effects.

The second challenge is transparency.

Many AI models operate as black boxes, making predictions that are difficult for scientists to interpret mechanistically. Regulators and pharmaceutical companies still need explainable evidence before advancing medicines into clinical development.

Third, expectations have become extraordinarily high.

Because AlphaFold achieved such a dramatic breakthrough, many investors expect Isomorphic Labs to revolutionize every stage of drug discovery immediately. In reality, translating computational advances into FDA-approved medicines will likely take years.

Finally, competition is accelerating.

Companies like Recursion, Insilico Medicine, Generate Biomedicines, Xaira Therapeutics, EvolutionaryScale, and numerous others are pursuing similar visions using different technological approaches.

The AI drug discovery race has only just begun.

Lessons from Isomorphic Labs

There are several lessons emerging from the Isomorphic story.

First, foundational scientific infrastructure often creates greater value than individual products. AlphaFold itself isn't a drug, but it fundamentally changed how biology is studied worldwide.

Second, interdisciplinary thinking matters. Isomorphic exists because computer scientists, structural biologists, physicists, mathematicians, and medicinal chemists learned to solve problems together.

Third, platform technologies compound over time. Every prediction improves future models, creating a positive feedback loop that traditional drug discovery struggles to replicate.

Finally, the future of pharmaceutical research may belong to organizations capable of integrating artificial intelligence with experimental biology rather than treating them as separate disciplines.

What's Next?

Today, Isomorphic Labs is moving beyond protein structure prediction toward designing complete therapeutic molecules.

The company continues expanding its AI foundation models while growing partnerships with major pharmaceutical companies. Industry observers expect additional collaborations, internally developed programs, and eventually its first clinical candidates.

Just earlier this year, the company entered into a strategic collaboration with Johnson & Johnson, further validating their platform’s cross-modality (small and large molecule) capability.

The company has scaled significantly (now with a US presence). Back in June of 2025, Dr. Ben Wolf was appointed Chief Medical Officer. He earned his B.S. from Union College and M.D. and Ph.D. in biochemistry from the University of Virginia, completing medical training in internal medicine and medical oncology at the University of California at San Diego. He also brings in more than 20 years of biopharma experience.

Looking toward 2030, the ambition extends even further.

Imagine AI systems capable of designing antibodies, small molecules, RNA therapeutics, protein degraders, or entirely new classes of medicines within weeks instead of years.

Imagine virtual laboratories where millions of experiments occur computationally before a single compound is synthesized.

That is the future Isomorphic Labs hopes to build.

Whether that vision becomes reality remains uncertain.

But regardless of the outcome, the company has already changed how scientists think about biology.

Outro: From Understanding Life to Engineering It

The story of Isomorphic Labs is not simply about artificial intelligence. It is about a fundamental shift in how science itself is conducted.

For generations, biology has largely been an observational discipline. Scientists collected data, formed hypotheses, and slowly uncovered nature's rules through painstaking experimentation. Isomorphic Labs represents a new paradigm—one in which artificial intelligence becomes an active participant in scientific discovery, capable of generating hypotheses, designing molecules, and accelerating research at a scale previously unimaginable.

Whether Isomorphic ultimately develops blockbuster medicines or simply becomes the platform powering future pharmaceutical breakthroughs, its influence is already undeniable. It has demonstrated that AI is no longer just a tool for analyzing biological data; it is becoming an engine for creating biology itself.

For the scientific translator, perhaps the biggest lesson is this: every era of medicine has been defined by a new technology—from chemistry to molecular biology, from genomics to gene editing. Artificial intelligence may become the next platform upon which the next century of drug discovery is built.

This has been Petri Dish Perspectives. I’m Manead. Thanks for listening. See you next Thursday. Good bye.

References

1. https://www.isomorphiclabs.com/
2. www.wikipedia.org
3. https://endpoints.news/ 
4. https://finance.yahoo.com/ 

© 2026 The Perspective Bureau LLC. All rights reserved.