Recursion Pharmaceuticals
"TechBio" pioneer that runs a scaled phenomics platform — automated cell imaging + ML — to discover drug candidates, now merged with Exscientia.
1. Core Product / Service
Recursion Pharmaceuticals (NASDAQ: RXRX) operates the Recursion OS, a platform that combines automated high-throughput cell imaging ("phenomics"), ML-driven feature extraction, and a proprietary biological dataset to predict drug candidates [1]. Its thesis, branded "TechBio", is that industrialized biological data plus machine learning can industrialize discovery the way "TechBio" differs from "BioTech".
After its November 2024 all-stock merger with Exscientia (~$688M), the combined company added Exscientia's precision chemistry / generative molecular design to Recursion's phenomics, giving it an end-to-end "design → screen" stack [1]. In June 2025 it also released Boltz-2, an open-source biomolecular foundation model built with MIT and NVIDIA [1].
2. Target Users & Pain Points
- Pharma partners (Roche/Genentech, Sanofi) licensing the platform or co-developing programs.
- Its own pipeline — Recursion runs internal oncology and rare-disease programs.
Pain solved: preclinical discovery is slow and failure-prone; Recursion positions industrial-scale data generation + ML to move failure earlier and cheaper (see ai-for-science).
3. Competitive Landscape
| Player | Positioning | Vs. Recursion |
|---|---|---|
| isomorphic-labs | AlphaFold-lineage, Alphabet-funded | Big-tech funded; Recursion is independent/public |
| insilico-medicine | Generative-AI-first biotech (Pharma.AI) | Recursion is phenomics/data-first |
| absci | De novo antibody design + wet lab | Antibody-focused; Recursion broader small-molecule/oncology |
| insitro | ML + causal biology (disease mechanism) | Insitro leans causal-biology discovery; Recursion leans scale |
4. Unique Observations
- Scale over cleverness: Recursion is the purest bet that industrialized wet-lab data generation — not a novel model architecture — is the moat in AI drug discovery; its automated imaging flywheel predates the generative-AI wave.
- "Trading at cash, period, if it weren't for AI": investor commentary this period framed most of Recursion's market value as resting on upcoming clinical readouts, underscoring how far the whole sector's valuation still hangs on proof rather than platform revenue [1].
- Consolidation as survival: the Exscientia merger is the sector's clearest signal that AI-biotech is consolidating around scale and cash runway before any blockbuster exists.
5. Financials / Funding
- IPO: 2021 (Nasdaq: RXRX) [1].
- NVIDIA strategic investment: $50M (July 2023) [1].
- Roche/Genentech: ~$150M upfront (Dec 2021); Sanofi collaboration has generated ~$130M inflows through Q1 2025 [1].
- Merger cash: ~$850M combined at Exscientia close (Nov 2024) [1].
- Q1 2025: revenue $15M; net loss $203M; cash $509M as of Mar 31, 2025 [2].
- 2025 restructuring:
20% workforce cut (160 roles) to extend runway to Q4 2027 [1].
6. People & Relationships
- Co-founder / CEO: Chris Gibson, PhD.
- Exscientia leadership: David Hallett (interim CEO) became CSO post-merger.
- Investors/partners: NVIDIA, Roche/Genentech, Sanofi.
- Peers: isomorphic-labs, insilico-medicine, insitro, absci.
Sources
- [1] Recursion Pharmaceuticals — AI Wiki profile (2026-08-24)
- [2] Nasdaq, "Recursion Reports First Quarter 2025 Financial Results…" (2026-08-24)
Recursion
自建表型成像平台的上市 AI 药企,向 Tempus 买数据并开源 RxRx
- 数据类型
- 高内涵表型成像;转录组
- 规模口径
- 最高 220 万实验/周(公司口径)
- 客户
- Sanofi;Roche;Genentech;Tempus
做什么
美国上市 AI 药物发现公司,以自动化表型成像为核心平台,开源 rxrx-datasets。
数据怎么生产
公司口径"最高可达每周 220 万次实验、1,620 万张多时间点明场图像(135 TB)"(FY2024 稿,"experiments"与"samples"措辞在官方页间不一);FY2025 稿改为"每周数百万次",未复现 2.2M。">50 PB"未在官方来源找到,不采用 [1]。
谁付费、数据流向
合作药企:Sanofi 已达成 $130M 上市前及里程碑,Roche/Genentech 累计 $213M(二手)。数据采购侧:向 tempus 买数据,2026-09-15 改为 3 年共 $4,200 万固定付款(第 3/4/5 周年各 $1,400 万,每期至少 $400 万现金,其余可用股票;原为两年内可能 $8,400 万的可选费用),同时将 RNA 模型 TxFM 非独占授权给 Tempus [2]。这是生物领域少有的公开现金买数据先例。
争议与口径
Boltz-2 评测使用 Recursion 内部 assay,但训练不含,见 boltz-2。TTM 收入约 $5,500 万为待复核数。