The $50B Question: How Do We Know If Isomorphic's AI Works?

Isomorphic Labs is in talks to raise about $5 billion at a valuation of about $50 billion, Bloomberg reported on November 4. The company has already raised about $2.7 billion from outside investors since 2025. Its president, Max Jaderberg, says its drug design engine works. But the company has not publicly named a drug candidate, and this week Jaderberg said its drug work is still preclinical.

A new manifesto... and skepticism

On September 29, Jaderberg published "Building a new path to make medicines with AI". "Our hypothesis has been proven," he writes, and the company's preclinical data is "helping us to gear up for clinical developments."

The centrepiece is a design run on an unnamed "Target D." Isomorphic's engine, IsoDDE, searched chemical space on its own for two to four days. A handful of molecules were then made and tested, and one produced a working concentration-response curve. Isomorphic compares that with a published molecule it estimates took three to five years of work.

Melinda B. Chu (@MelindaBChu1) was unimpressed. "Isomorphic is slower than traditional R&D and with $2 billion," she wrote. "No IND in 5 years. Not even real preclinical data reported; only vague Target D." Her prediction: no trial in 2026, and possibly not in 2027.

We replied that counting drug candidates might not be the only measure of AI success. She pushed back. Isomorphic is a drug company, she said, its founders talk about curing disease, and "you can't cure a disease without a drug." She also called our view "overly optimistic."

Investors seem to be siding with the optimists, at least for now. This piece sets out both cases and the public record behind each.

The skeptics' case: big money, delays, little disclosure

The money is real, and growing fast. Isomorphic raised $600 million in its first outside round in 2025. In May 2026 it raised a $2.1 billion Series B, led by Thrive Capital with Alphabet, GV, Temasek and others. That is about $2.7 billion from outside investors, on top of Alphabet's earlier funding. In October, Bloomberg reported early talks for a new round valuing the company at $40 billion or more. By early November, the talks were for about $5 billion at about $50 billion. That round has not closed, and terms could change. If it completes, outside funding would approach $8 billion.

The clinic keeps moving. Demis Hassabis had said Isomorphic would have AI-designed drugs in trials by the end of 2025. At Davos in January 2026, he pushed that to the end of 2026. In mid-September, head of drug discovery Chris Butler said the company was "on track" in preclinical work but declined to give a clinical timeline. On November 2, at Bloomberg Tech in London, Jaderberg said the drug work was still in preclinical stages and described the company as in its "first innings."

Almost nothing is disclosed. There is no named candidate, no named target and no program count. The only pipeline detail is two focus areas: oncology and immunology. Target D has a curve but no potency figure, no selectivity or ADMET data and no in vivo results.

Some AI peers have arguably moved faster. Chu points to Insilico Medicine, which took rentosertib, its drug for idiopathic pulmonary fibrosis, from target discovery to phase 1 in under 30 months. It has since published phase 2a results in Nature Medicine and started phase 3. The field is also getting more crowded. OpenAI released a life sciences model this year, and OpenAI-backed Chai Discovery raised $400 million in July.

As of early November, Isomorphic has not announced an IND filing or a trial start.

The optimists' case: partners, benchmarks, investors and normal biology timelines

The case for Isomorphic rests on partnerships, benchmarks, investor conviction, and timelines.

Big pharma keeps signing, and coming back. In January 2024, Isomorphic signed its first two deals. Eli Lilly paid $45 million upfront, with up to $1.7 billion in milestones; Novartis paid $37.5 million upfront, with up to $1.2 billion. A year later, Novartis added up to three more programs on the same terms. In January 2026, Johnson & Johnson joined with a multi-target deal covering several types of molecule; terms were not disclosed. Pharma companies that have seen the internal data are expanding, not leaving.

The engine has public numbers. In February 2026, Isomorphic published a technical report on IsoDDE. It claims more than double AlphaFold 3's accuracy on hard protein-ligand cases, better antibody-antigen prediction and binding-affinity estimates that match physics-based methods. These are benchmarks, not drugs, but they can be checked.

Investors keep raising the price. Isomorphic's valuation has climbed in the space of a month without any new public data, from a reported $40 billion or more in October to about $50 billion in the latest talks. Whatever investors have seen, it has not put them off.

Five years without an IND is not unusual. The industry average from a first hit to a candidate molecule is 33 to 36 months, before IND-enabling safety studies even begin. Isomorphic also spent its early years building the platform, including AlphaFold 3 with Google DeepMind in 2024, rather than running programs from day one.

What is a $50 billion platform worth without a drug?

In our exchange, we pointed to CytoReason, which sells AI disease models to pharma and has never built a pipeline. Pfizer extended its deal with CytoReason in 2022 for up to $110 million over five years; the company has raised about $130 million in total. Chu replied that Isomorphic was never pitched as a software company, and that platform revenue would not come close to its valuation.

Schrödinger, which sells drug design software and also runs its own pipeline, gives a sense of scale. In 2025 it reported $199.5 million in software revenue and a $103 million net loss. Isomorphic does not sell its engine at all. Its income comes from partners, and the public figures are $82.5 million in upfront payments from Lilly and Novartis. A $50 billion valuation would be roughly 600 times that amount. We found no public report of a milestone payment, which would show a partner program reaching an agreed scientific goal.

Isomorphic set its own test: a first clinical trial by the end of 2026, itself a delay from the earlier goal of late 2025. With weeks left in the year and its president describing the work as preclinical, Chu's prediction looks increasingly likely to hold. Investors, though, are pricing the company at tens of billions anyway. If the trial slips into 2027, will that say something definitive about the platform's limits, or only that biology takes time? For now, the market has made its bet before any drug has reached a patient.

Go deeper into the increasingly lucrative and sophisticated techbio scene with BiopharmaTrend Pro Membership.

Topic: AI in Bio