|
Getting your Trinity Audio player ready...
|
AstraZeneca has handed a California biotech a chance at one of the larger artificial intelligence drug discovery deals of the year, but the real money depends on science that has not been done yet. The British drugmaker announced on Thursday, October 15, a multi-year collaboration with insitro to find and develop treatments for metabolic dysfunction-associated steatohepatitis (MASH), a serious fatty liver disease, and for cardiomyopathies, diseases of the heart muscle. Insitro will receive $40 million upfront and is eligible for up to about $1.46 billion in further milestone payments, bringing the headline value to roughly $1.5 billion, plus tiered royalties.
The structure is typical of biopharma partnerships and worth reading closely. Of the total, only the $40 million is guaranteed, which is under 3 percent of the headline figure by simple arithmetic. The rest is conditional on insitro hitting discovery, development and commercial milestones, and the company would also earn single-digit percentage royalties on global sales of any products that reach the market, according to the companies’ announcement and coverage by Benzinga and Seeking Alpha. Most of the deal’s headline value depends on later development and sales milestones, as Runtime Wire put it, so the agreement will test whether insitro can turn machine-learning-led target discovery into programs that AstraZeneca chooses to advance. One outlet’s headline cited a larger $1.86 billion figure for the same deal, which does not match the companies’ own numbers, so the $1.5 billion figure should be treated as the reliable one.
The science behind the pact is what AstraZeneca is paying for. Insitro will use its Virtual Human platform, which combines lab-grown human cell models with machine learning, to identify the causal biology of disease, according to Traders Union’s summary. Insitro will research targets nominated by AstraZeneca and advance them from discovery to the development candidate stage, meaning the point at which a specific molecule is chosen for early testing. A drug target is a molecule in the body, often a protein, that a medicine is designed to hit. Picking the wrong one is one of the biggest reasons that drug programs fail, which explains why AstraZeneca’s executives emphasise it.
Sharon Barr, AstraZeneca’s executive vice president of biopharmaceuticals research and development, said selecting new targets for the pipeline is among the most difficult parts of drug discovery and development, according to Benzinga. She expects the partnership to improve target selection and speed up the search for potential therapies, and said that combining AI tools with expert scientific work reduces the number of molecules that must be tested in the laboratory, decreases uncertainty and accelerates the hunt for promising drugs. Those are statements of expectation, not results, and the reports reviewed contain no data showing that the approach will work for these diseases.
The two diseases are well chosen for a data-driven approach. MASH, previously known as non-alcoholic steatohepatitis, is a form of fatty liver disease that can progress to scarring and liver failure, and cardiomyopathies are a group of conditions in which the heart muscle weakens or thickens. Both are described as areas with significant unmet medical need, according to Seeking Alpha. They are also biologically complex, with many interacting causes, which is where large-scale human and cellular data analysis is supposed to help. The platform combines large human and cellular datasets with AI analysis to identify promising targets and support the design of new medicines, Benzinga reported.
Insitro’s founder and chief executive, Daphne Koller, a Stanford-trained computer scientist, has focused the company on early-stage discovery, according to a profile summary of the deal. The company has been broadening its list of pharmaceutical partners, including a collaboration with Bristol Myers Squibb on ALS research and one with Eli Lilly on machine-learning models for predicting how molecules behave in the body. AstraZeneca joins that list, and the reports reviewed do not give financial terms for the earlier partnerships, so it is not possible to compare deal sizes.
For AstraZeneca, the agreement fits a broader dealmaking push. Seeking Alpha noted that the company has stepped up its activity as it targets $80 billion in total revenue and 20 new medicines by 2030. eMarketer reported that since late 2024, AstraZeneca has licensed cholesterol, obesity and AI-driven discovery programs from China’s CSPC, including a $1.2 billion upfront payment in January, and that it bought the AI company Modella in January to bring its platform and agents into oncology drug development. The insitro deal adds an earlier-stage approach to that mix, focusing on finding targets in the first place.
The wider context is a boom in pharma-AI partnerships. DealForma data cited by eMarketer shows AI-based target and drug discovery licensing deals reached $47.1 billion in the first half of 2026, already above the $44.6 billion recorded for all of 2025. Those totals are headline values that include contingent payments, which means the money actually changing hands is a fraction of the figure. IQVIA’s May 2026 research and development trends report, cited in the same piece, found that AI use in early target identification remains much less common than in later stages of drug discovery. The surge in deals puts more pressure on companies to demonstrate results, eMarketer noted, since AI target discovery offers a way to build long-term pipelines but does not address immediate revenue pressures.
It is important to keep expectations realistic. Drug development is slow, expensive and high-risk: even good targets often fail in clinical trials, and a candidate that emerges from this collaboration would still need years of testing before it could reach patients. Nothing in the announcement says a medicine is close, and the earliest outcome is the identification of targets and development candidates, which are preliminary steps. The deal is therefore best read as a bet that AI can improve the odds at the start of the process, not as evidence that it already has. Skeptics will point out that the industry has announced many AI partnerships with large headline values and that few have yet produced approved medicines, though the reports reviewed do not give a count.
For investors and the public, the measures to watch are specific. Milestone payments, which companies sometimes disclose in later filings, will show whether insitro’s programs are progressing. Any naming of targets or candidates, and the start of early-stage studies, would be concrete signs. The sector as a whole will be judged on whether AI-derived programs reach and succeed in clinical trials, not on the size of the deals that precede them. AstraZeneca’s own announcements are published on the AstraZeneca website, and insitro describes its platform on its insitro site.
Readers who follow global technology, health and business stories can find more coverage at BusinessTech Nigeria. The deal is a clear statement of intent from a major drugmaker, and its value will become clear only if the targets it uncovers turn into medicines for patients with liver and heart disease.
