Demis Hassabis, and how AI just might wrangle our molecular universe

Isomorphic Labs, founded by Demis Hassabis in 2021, is positioning advanced AI to tackle drug discovery with an ambitious mission to “solve disease,” but the company currently has no assets in clinical trials and emphasizes the need to demonstrate validated, patient-facing results. The piece also catalogs sizable private-market activity: Mews raised $300M (Series D) led by EQT Growth; Inferact and Preply each raised $150M; LiveKit and Railway raised $100M apiece; EQT agreed to acquire Coller Capital for up to $3.7B; and BitGo plans an IPO to raise up to $200.6M (11.8M shares at $15–$17).

Analysis

Market structure: Winners are platform owners (Alphabet/GOOG, Schrodinger/SDGR, large CROs) and private-equity consolidators (EQT) that capture recurring software/outsourcing economics; losers are asset-heavy, preclinical biotechs that rely on legacy R&D models and firms without large proprietary datasets. Competitive dynamics will favor players that sell validated, repeatable workflows (pricing power for SaaS-style models) while atomized discovery shops face margin pressure and acquisition or failure risk. Cross-asset: a credible AI-drug validation would re-rate growth equities, tighten high-yield spreads for biopharma financing and marginally strengthen USD via risk-on flows; commodities impact is negligible.

Risk assessment: Tail risks include failed pivotal trials for AI-derived candidates, regulatory limits on model explainability, IP litigation between incumbents and AI startups, and a macro funding pullback; any single failure could knock 30–60% off exposed small-cap names. Time horizons: sentiment moves in days-weeks around funding/IPO headlines, medium-term (3–12 months) depends on clinical readouts and partner deals, while structural effects play out over 1–3 years as one or two success stories emerge. Hidden dependencies: training-data access, CRO partnerships, and talent concentration create single points of failure. Key catalysts: Isomorphic/DeepMind disclosures, EQT–Coller close, Alphabet earnings commentary and first clinical candidate disclosures in the next 6–18 months.

Trade implications: Favor 12–18 month, concentrated exposure to GOOG (buy LEAPS) and EQT (buy stock or call spread) while trimming pure-play small-cap AI-discovery names with <12 months runway. Implement pair trades: long SDGR (platform/software) vs short RXRX (recursion-style high-burn discovery) sized to 1–2% NAV each, and use 6–12 month put spreads on the short leg to limit tail risk. Sector rotation: increase software & CRO weights by +3–5% and reduce speculative biotech exposure by -3–5%. Entry/exit: scale in over 4–8 weeks ahead of Alphabet earnings and EQT deal close; trim positions by 30–50% on any negative clinical readout or by 12 months if no clinical candidate advances.

Contrarian angles: The market underestimates the multi-year path from molecule design to approved drug — don’t pay for promise alone; current investor enthusiasm likely overprices small public discovery names while underpricing big-tech optionality (GOOG, MSFT). Historical parallel: genomics/software hype cycles where infrastructure/software winners outlast wet-lab outsiders — expect consolidation, not broad-based winners. Unintended consequences: increased litigation/IP disputes and slower regulator approvals could make private assets acquisition targets for PE (EQT) at discounted multiples, creating asymmetric entry points in 12–24 months.

Key Decisions for Investors

• Establish a 2–3% portfolio position in Alphabet (GOOG) using 12–18 month LEAPS (target strike ~+15% from spot) to capture platform upside; scale in over 4–8 weeks before Alphabet earnings and reassess on any DeepMind/Isomorphic disclosures.
• Take a 1–2% long in EQT (EQT) stock or a 12-month call spread (limit debit to <1% NAV) ahead of the Coller close, targeting a 20–40% upside if integration synergies materialize within 12 months.
• Initiate a relative-value pair: long Schrodinger (SDGR) 1–2% NAV vs short Recursion (RXRX) 1–2% NAV (or similar high-burn AI-discovery small caps); hedge the short with 6–12 month put spreads to cap downside and target 25–50% relative outperformance over 9–18 months.
• Reduce pure-play speculative biotech exposure by 3–5% of portfolio weight; redeploy into CROs and software/SaaS names serving drug discovery (increase weights by +3–5%), expecting clearer revenue paths in 6–18 months.
• Avoid enlarging positions in public AI-drug wannabes with <12 months cash runway; if considering, require upcoming clinical or partnership catalyst within 90 days or a cash runway >18 months before allocating capital.