Isomorphic Labs

Type: Private AI drug design company — DeepMind spinout

Isomorphic Labs was spun out of Google DeepMind in 2021 by Demis Hassabis, built on the AlphaFold protein structure prediction platform. The thesis: if you can predict protein structures perfectly (AlphaFold), you have the foundation for rational drug design — using the structure to design small molecules and biologics that interact precisely with the target.

Why it's the sector's most-watched company: Isomorphic sits at the intersection of (a) the best foundation model in biology (AlphaFold), (b) the most credible founder (Demis Hassabis, Nobel Prize co-laureate 2024 with David Baker for protein design), and (c) the most ambitious goal (first fully AI-native drug company to take a molecule from AI design to Phase 1). If it hits its IND target by end-2026, it becomes the most important proof point for the techbio thesis.

Funding

• Series B (January 2026): $2.1B raised from Thrive Capital, MGX (Abu Dhabi), Temasek, UK Sovereign AI Fund
• Total raised: ~$2.6B
• Source: IntuitionLabs

Pharma Partnerships (deal sizes are upfront + milestone totals)

• Eli Lilly: Up to $1.7B (announced January 2024)
• Novartis: Up to $1.2B (announced January 2024)
• Pre-revenue contracted value: >$2B across named partners

Clinical target

• First internal IND submission: targeted by end-2026
• If achieved: would be the first AI-native company to take a fully internally-discovered drug into Phase 1 from its own pipeline (distinct from AI-assisted discovery at pharma partners)

IsoDDE Engine & Pipeline (technical report, 2026-02-10)

• IsoDDE (Isomorphic Drug Design Engine): on the hardest protein-ligand structure-prediction cases, IsoDDE achieved 50% accuracy vs AlphaFold 3's 23.3% — more than double the system underpinning the 2024 Nobel-recognized work.
• 17 active programs across oncology, immunology, and cardiovascular.
• First AI-designed drugs to enter Phase 1 by end of 2026 (including, per reporting, the first AI-designed cancer drug).
• Architecture stance: platform-thesis flagship — engine first (IsoDDE), assets second.

Key Claims

• IsoDDE roughly doubles AlphaFold-3's accuracy on the hardest protein-ligand cases (50% vs 23.3%) — Per Isomorphic's IsoDDE technical report (2026-02-10), corroborated across multiple 2026 reports. Evidence: moderate (Isomorphic IsoDDE)
• 17 active programs; first AI-designed drugs targeted for Phase 1 by end-2026 — Spanning oncology, immunology, and cardiovascular. Evidence: moderate (Isomorphic IsoDDE)
• Isomorphic is the platform-thesis flagship, but the platform→asset bridge is unproven — AI drug discovery has drawn more capital since 2024 than any biotech subsector since immuno-oncology, yet has zero approved drugs to date. Evidence: moderate (Isomorphic IsoDDE)

Provenance: the IsoDDE source file is summary-derived (host returned HTTP 403), corroborated across multiple 2026 reports; underlying primary = Isomorphic's IsoDDE technical report (2026-02-10).

What to watch

• End-2026 IND announcement (or slip)
• First Phase 1 safety readout (~2027–2028)
• Whether AlphaFold-derived design advantage translates to better binding affinity in clinical candidates vs traditional structure-based design

Sources

• Isomorphic Labs $2.1B Series B analysis (IntuitionLabs)
• letsdatascience.com on Series B
• Isomorphic IsoDDE + first human trials (2026) — summary-derived; primary = IsoDDE technical report 2026-02-10

Changelog

• 2026-06-24 — Compiled new sources (isomorphic-isodde-human-trials)