insitro discovers drug targets and designs experimental medicines for its own pipeline and pharmaceutical partners, focusing on metabolic disease and neuroscience. Its product for a pharmaceutical research team is a validated biological target or drug candidate, rather than a general-purpose chatbot for scientists.

For a liver-disease program, scientists start with human genetic, imaging, and clinical data to identify biological changes associated with liver fat or fibrosis. insitro's Virtual Human platform uses machine learning to extract measurable disease characteristics and connect them to genetic variation, helping scientists prioritize genes that may cause disease rather than merely accompany it. Researchers then test whether changing a selected gene alters disease-relevant behavior in human cells, before advancing a potential treatment into animal studies and development work required for human trials.

The IRS1 program illustrates this workflow: analysis incorporating UK Biobank data identified a liver-fat target, and insitro designed CTRO-1013, an investigational siRNA that reduces expression of that gene in the liver. Reported effects in human hepatocytes and animal models support continued development; they do not establish effectiveness in patients. In October 2026, insitro said its first liver programs were expected to enter the clinic in early 2027, within a nine-program pipeline.

Its TherML engine supports the molecule-design step across small molecules, oligonucleotides, and biologics, considering binding activity alongside properties such as stability and manufacturability. The announced January 2026 acquisition of CombinAbleAI added antibody-design expertise to this effort. For neuroscience, insitro has built more than 200 engineered and patient-derived ALS cell lines, allowing researchers to test changes in human motor neurons and identify targets for Bristol Myers Squibb.

insitro combines pharmaceutical discovery partnerships with development of medicines whose commercial rights it retains. Both use the same research infrastructure, but allocate development costs and potential drug-sales proceeds differently.

Pharmaceutical companies pay insitro to discover targets and advance selected programs, with compensation tied to upfront agreements, research progress, development milestones, and eventual sales. In its October 2026 AstraZeneca agreement, insitro is eligible for approximately $1.46B in potential milestones beyond the upfront payment, plus single-digit royalties on commercial sales; AstraZeneca takes responsibility for clinical development, regulatory work, and commercialization. These potential payments depend on future achievements and are not current revenue.

Automated laboratories, specialist researchers, biological assays, data access, and compute are inputs to this business. Reusing experimental systems and models across programs could spread those costs, while partnering transfers later clinical-development spending to a larger pharmaceutical company. Repeat target selections and extensions of the Bristol Myers Squibb collaboration illustrate how successful research can expand an existing relationship.

insitro's 2024 metabolic-disease agreements with Lilly run in the other direction: insitro can license Lilly's liver-delivery technology and collaborate on antibody development while retaining global rights to its programs. Lilly is eligible for milestones and royalties, and insitro assumes remaining development and commercialization after candidate nomination. Retaining rights creates exposure to future drug-sales economics, but also leaves insitro responsible for the cost and risk of developing those medicines.

insitro competes with Recursion, Xaira Therapeutics, and Isomorphic Labs for pharmaceutical partnerships, scientific talent, and ownership of valuable drug candidates. The contest spans both finding the right biological target and designing a medicine against it: predictive models are commercially useful only if they produce candidates partners will fund and eventually demonstrate safety and efficacy.

Recursion combines industrial-scale cellular experiments with multimodal data and has programs in human trials. insitro similarly invests in laboratory-generated data, but emphasizes connecting cellular findings to human genetics and clinical observations before choosing targets; its clinical results will determine whether that approach improves drug development rather than simply generating additional hypotheses.

Xaira is a particularly close competitor because it also combines cellular perturbation data, target discovery, and AI-designed therapeutics. Its antibody-design capabilities overlap with insitro's expansion into biologics, narrowing the distinction between their integrated discovery platforms and increasing the importance of individual drug results.

Isomorphic Labs builds on AlphaFold-derived structural modeling and has partnerships with Novartis, Lilly, and Johnson & Johnson, giving it established pharmaceutical distribution. insitro's relative proposition is to integrate disease-target discovery with molecule design, supported by its human and cellular datasets.

insitro can reach a different customer base by distributing individual predictive models beyond its bilateral agreements with large pharmaceutical companies. Smaller biotechnology teams also need to assess drug candidates, but may lack the proprietary experimental histories needed to train equivalent models.

In October 2026, models insitro built with Lilly became available to certain biotech customers through Lilly TuneLab. Trained on Lilly's historical data, they predict how small molecules behave in mice, rats, dogs, and monkeys, helping researchers prioritize compounds before running animal studies. This creates a route into the research budgets of companies that are not insitro discovery partners, although public announcements do not establish a separate subscription business or disclose insitro's economics from that distribution.

Clinical translation: insitro's liver programs are approaching first-in-human trials, with early 2027 the latest announced timeline. Genetic associations and improvements in cell or animal models may not translate into safe, effective treatments, jeopardizing both owned assets and the platform's appeal to partners.

Partner-dependent cash flows: Disclosed collaboration payments are concentrated in a small group of pharmaceutical partners and depend partly on research milestones. A partner that changes priorities or declines further targets can reduce incoming cash even while insitro continues funding its laboratory infrastructure.

Cost of retaining drug rights: The Lilly agreements leave insitro with global rights but responsibility for later development and commercialization. Advancing owned medicines alongside a broad discovery platform could require substantial additional capital before successful clinical results or drug sales provide a return.