Company

insitro

Daphne Koller's machine-learning-first drug company — discovers disease mechanisms from causal biology rather than screening existing compounds.

1. Core Product / Service

insitro (the name merges in silico + in vitro) applies machine learning to human disease biology, focused on discovering disease mechanisms — causal biological drivers — rather than screening known compounds [1]. Its platform pairs high-throughput biological data generation with ML: the Virtual Human™ engine for causal biology discovery and ChemML™ for small-molecule design [1]. Therapeutic focus spans metabolic disease, neuroscience, and oncology.

Its partnerships are the business model: Gilead (NASH), Bristol Myers Squibb (ALS), and Eli Lilly (metabolic / ML models for ADMET), with insitro in several deals retaining the clinical-development role [1][2].

2. Target Users & Pain Points

  • Pharma partners who license validated targets or co-develop programs.
  • Its own pipeline — insitro is building toward in-house candidates, with a first clinical-trial entry targeted for 2026 [1].

Pain solved: target selection is the highest-leverage failure point in drug R&D; insitro's causal-biology approach aims to pick better targets before chemistry begins, versus optimizing later stages.

3. Competitive Landscape

Player Positioning Vs. insitro
recursion-pharmaceuticals Phenomics + ML at scale Recursion is data-scale screening; insitro is causal-mechanism discovery
isomorphic-labs AlphaFold lineage, Alphabet Structure-prediction anchor vs. disease-biology anchor
insilico-medicine Pharma.AI generative design Insilico is generative-chemistry-first; insitro is mechanism-first

4. Unique Observations

  • Founder as brand: Daphne Koller (Stanford professor, co-founder of Coursera, ex-Calico chief computing officer) gives insitro the strongest academic-AI pedigree in the sector — the "ML pioneer builds a biotech" archetype [1].
  • Mechanism-first vs. molecule-first: insitro is the cleanest test of the thesis that AI's biggest drug-discovery leverage is earlier — in target/mechanism selection — whereas most peers (absci, chai-discovery, insilico-medicine) apply AI later to molecule/antibody design.
  • The "eight years to a first trial" timeline: insitro's ~2026 first clinical entry, eight years after founding, is a reminder of how long even the best-capitalized AI-biotechs take to reach the clinic — counterweight to sector hype [1].

5. Financials / Funding

  • Rounds: $100M Series A (2018), $100M Series A+ (2019), $143M Series B (2020), $400M Series C (2021, led CPP Investments with a16z, BlackRock, Arch, GV, Third Rock) — total $743M ($800M per company) [1].
  • Partnerships (potential value): Gilead (>$1B, $15M upfront), BMS (ALS, $50M upfront, >$2B potential), Lilly — combined deal value >$5B, ~$150M collaboration revenue received [1][2].
  • Valuation: estimated $2.6–3.2B (2025) [1].

6. People & Relationships

Sources

  • [1] insitro — Lilly partnership announcement (2026-08-24)
  • [2] insitro — BMS collaboration expansion (2026-08-24)
Data profile · AI for Science

insitro

机器学习加高通量细胞实验的药物发现公司,公开材料有限

现有信息

insitro 是美国的 ML 驱动药物发现公司,按公开信息归入自建实验数据路线,具体实验平台细节未核实 ⚠️。公开渠道未取得其数据规模、融资、客户与合同条款的一手来源 ⚠️。

定价与外部数据

WuXi、Insitro 的公开服务价格未查到。是否向外部出售数据或提供按次服务,未核到。

口径提示

同类自建数据工厂的通量多为公司口径(如 recursion-pharmaceuticals、Absci 自报约 280 万设计/周),缺独立验证。对比时宜写"未披露"而非估算。补查方向:数据规模、合作中的数据归属条款,参见 data-ownership-terms。

  1. https://www.1stoncology.com/blog/?p=108939 (2026-10-15)
Last compiled: 2026-08-24