Isomorphic Labs Closes $2.1B for AI-Driven Drug Discovery

At a glance

• $2.1 billion Series B closed on 12 May 2026 — one of the largest applied-AI rounds of the year.
• Thrive Capital leads the round, joining a syndicate that includes Alphabet, GV, MGX, Temasek, CapitalG and the UK Sovereign AI Fund.
• Use of proceeds: scale the IsoDDE drug-design engine, advance internal therapeutic candidates toward the clinic, expand partnered programmes, and hire across AI, engineering, drug design and clinical disciplines.
• Three sites — London (HQ), Cambridge Massachusetts and Lausanne — anchoring a transatlantic applied-AI biotech footprint.
• Builder takeaway: vertical AI for regulated sectors is no longer a side bet. The capital is flowing toward applied frontier-AI in pharma, biology and defence.

Inside the $2.1B round

Isomorphic Labs confirmed on 12 May 2026 that it had closed a $2.1 billion Series B round — roughly £1.65 billion at prevailing rates — making it one of the most heavily capitalised vertical-AI companies on the planet. The DeepMind spinout was founded in 2021 by Sir Demis Hassabis, the Nobel-laureate co-founder of DeepMind, with a single mandate: turn the AlphaFold breakthrough into an industrial pipeline of drug candidates.

The cap table for this round splits cleanly into two camps. On one side sit the new growth investors. Thrive Capital, Joshua Kushner's New York firm that has become a fixture in late-stage AI rounds, leads. They are joined by MGX, the Abu Dhabi state-backed AI vehicle that has rapidly become a fixture across frontier-AI cap tables, and by Temasek, Singapore's long-horizon sovereign investor. Alphabet's later-stage growth arm CapitalG also wrote a new cheque, distinct from its earlier-stage sibling.

On the other side sit the existing backers doubling down. Alphabet itself participated again, as did GV, Alphabet's early-stage venture arm. The presence of three Alphabet-linked entities — Alphabet, GV and CapitalG — across one Series B is unusual and signals how strategically Mountain View is treating drug discovery as an applied-AI category.

And then there is the politically interesting cheque: the UK Sovereign AI Fund. Its participation puts a national-strategic stamp on the round and is the clearest signal yet that Britain's £500 million applied-AI vehicle intends to back champions in regulated sectors, not only foundation-model labs. For context on how that fund is being deployed across the wider portfolio, see our coverage of the UK Sovereign AI Fund's first investments.

What this cap table tells us, in plain English: Isomorphic is being treated like an applied-AI national champion by the UK, like a strategic AI bet by Alphabet, and like a generational growth opportunity by global crossover capital. That is a rare three-way alignment.

What the IsoDDE engine actually does

IsoDDE — the AI Drug Design Engine — is the system Isomorphic Labs has been quietly building since spin-out. It is the descendant, in spirit and in some of its lineage, of AlphaFold: the model that solved the half-century-old protein-folding problem and earned the 2024 Nobel Prize in Chemistry for Hassabis and John Jumper. But IsoDDE is not AlphaFold. AlphaFold predicts structure. IsoDDE designs molecules.

The practical workflow looks roughly like this. A target is identified — a protein implicated in a disease. IsoDDE generates candidate small molecules that should bind to the target, predicts their binding affinities, scores them for drug-like properties (absorption, distribution, metabolism, excretion and toxicity), and proposes a ranked shortlist for wet-lab validation. Each cycle of synthesis and assay feeds the next round of generation. The promise is dramatic: collapse the historical four-to-six-year early-discovery timeline into something measured in months.

This is the same applied-AI playbook that DeepMind has been refining elsewhere in its alumni network. Our coverage of AlphaEvolve, DeepMind's algorithm-discovery agent, shows the pattern: take a frontier-AI capability, point it at a hard scientific problem, close the loop with empirical validation. Isomorphic is doing the same thing in chemistry.

The $2.1 billion is the cost of moving from "compelling demos" to "drugs in patients". That transition requires three things at scale: compute to train and run ever-larger structure-and-function models; wet-lab infrastructure to validate the AI-generated hypotheses; and clinical operations to push the best candidates into trials. Isomorphic is now resourced to do all three concurrently, which is what separates a research effort from a pharmaceutical company.

UK Sovereign AI Fund Year 1: a portfolio view

The UK Sovereign AI Fund's appearance on Isomorphic's cap table is part of a busier Year 1 than many observers expected. The fund has now backed three high-profile rounds in quick succession.

First came Ineffable Intelligence's $1.1 billion seed, the David Silver-founded London lab pursuing a frontier-model agenda directly. Second was Recursive Superintelligence's $500M-plus round, oriented around self-improving systems. Now Isomorphic — a vertical applied-AI bet in pharma — completes a clearly intentional triad.

Read together, the portfolio reveals a thesis. The UK is hedging across three layers of the applied-AI stack: a horizontal frontier-model lab, a horizontal self-improvement research bet, and a vertical applied-AI champion in a sector where Britain already has world-class talent and infrastructure (pharma). It is a smarter portfolio than simply funding three more foundation-model labs would have been, and it leans into Britain's actual industrial strengths.

This is also consistent with the broader UK talent picture. As our analysis of 112 DeepMind alumni startups showed, the country is now experiencing a once-in-a-generation cluster effect — DeepMind veterans spinning out into a dense ecosystem of applied-AI companies. Isomorphic is the largest, but it is far from alone, and public capital is finally moving in step with that talent wave. Compare this to UKRI's £40m Fundamental AI Research Lab, which targets the upstream basic-science layer. The two together — sovereign fund downstream, UKRI lab upstream — start to look like a coherent national strategy rather than a series of disconnected announcements.

What IN + UK applied-AI builders should take away

For founders, engineers and operators in India and the UK, this round carries five practical signals.

1. Vertical AI is now where the biggest cheques live. If you are an Indian builder weighing whether to chase a horizontal LLM wrapper or a deep vertical (clinical workflows, agritech, energy modelling, materials science, manufacturing), the capital is unmistakably moving toward the vertical end of the spectrum. India has analogues in the making: IISc and IIT Madras both have credible computational-biology programmes; AI4Bharat's multilingual stack at IIT Madras is a clean precedent for sector-focused applied-AI built in-country. The lesson is not "copy Isomorphic" but "build the IndiaAI Mission's equivalent for a sector where India has structural advantages".
2. Pharma capital is reachable for AI teams in the UK. Cambridge, Oxford and London together host one of the densest life-sciences clusters in the world. UK applied-AI builders who can credibly bridge the AI-and-wet-lab gap are now sitting on top of a stack of patient capital that simply did not exist two years ago. The UK Sovereign AI Fund's cheque to Isomorphic confirms that policy is finally aligned with the country's biology-and-AI thesis.
3. Three-city footprints are the new default. London-Cambridge MA-Lausanne is a deliberate choice. London anchors AI and capital; Cambridge MA anchors biotech talent and FDA proximity; Lausanne anchors European chemistry and the EPFL spinout pipeline. Indian builders should think about analogous triads — Bengaluru-London-Singapore is the obvious one for fintech-AI; Bengaluru-Boston-London works for healthtech-AI; Hyderabad-Cambridge-San Francisco maps to enterprise SaaS-AI.
4. Hiring premia are about to compound for hybrid talent. Isomorphic is hiring across AI, engineering, drug design and clinical roles. The talent that commands the biggest premium in 2026 is the engineer who can read a clinical protocol, or the biochemist who can write CUDA. India and the UK both have this hybrid talent in real numbers — but mostly untapped. If you are running an applied-AI team, your retention strategy needs to reward the bilingual engineer disproportionately, before Isomorphic, Recursion and their peers do.
5. The IndiaAI Mission needs a vertical-AI lane. The current IndiaAI Mission, even with its proposed budget doubling, has focused heavily on sovereign foundation models, GPU capacity and the Bharat-language compute stack. None of those is wrong. But the Isomorphic round shows what is missing: a deliberate vertical-AI lane that funds applied-AI champions in pharma, agriculture, climate and manufacturing. The UK Sovereign AI Fund's portfolio is a fair template — and one India would do well to adapt before the talent migrates.

What to watch next

Three threads to keep an eye on over the next four quarters.

One — the first IsoDDE-originated clinical asset. Isomorphic's existing partnerships with Novartis and Eli Lilly date back to 2024 and have produced disclosed pre-clinical candidates. The market test is which IsoDDE-originated molecule enters human trials first, and how its preclinical efficacy compares to traditionally discovered candidates. If even one IsoDDE molecule clears Phase I within 18 months, the entire AI-drug-discovery category re-rates.

Two — UK Sovereign AI Fund Year 2. If Year 1 funded a foundation-model lab, a self-improvement bet and a vertical-AI champion, what comes next? Defence-AI is the obvious gap, and the consultative noises from Whitehall suggest it is coming. The fund will likely also begin its first secondary purchases of equity from earlier-stage UK-AI startups, which would be an important signal for the broader cap-table economics of British AI.

Three — Indian biotech-AI follow-through. Does IndiaAI Mission, IISc or one of the IIT institutional venture vehicles produce a domestic Isomorphic equivalent? The talent and the data are there. The wet-lab and clinical operations infrastructure is the bottleneck. Whichever Indian institution solves that bottleneck first will define the country's applied-AI biotech identity for the decade.

Sources for this story: Isomorphic Labs press release; PR Newswire; TechFundingNews; BioPharm International; Fierce Biotech.