According to Zhitong Financial News, AstraZeneca (AZN.US) has announced a collaboration with machine learning biotech company Insitro, with a total potential value of up to $1.86 billion, aimed at helping the UK-based pharmaceutical company discover new drugs for liver and heart diseases. On Thursday, AstraZeneca stated it would pay an upfront fee of $40 million. Insitro will combine its own data with AstraZeneca’s data to better understand the causes of fatty liver disease and cardiomyopathy. If this approach proves successful, the biotech company could receive additional payments of up to $1.46 billion, along with potential sales royalties.
This transaction is the latest bet by the pharmaceutical industry that AI can accelerate the discovery of new drug targets. Insitro has previously announced R&D partnerships with Eli Lilly (LLY.US) and Bristol Myers Squibb (BMY.US). The California-based biotech company is backed by investors including Andreessen Horowitz (a16z), BlackRock, and Temasek.
Insitro uses its AI platform to simultaneously analyze clinical data and disease models, aiming to gain novel insights into the underlying causes of diseases and thereby increase the likelihood of successful treatments.
AstraZeneca's ambitions
“Choosing our new target is one of the hardest things we do,” said Sharon Bar, Executive Vice President and Head of BioPharmaceuticals R&D at AstraZeneca.
Barr said in an interview. “This is a real opportunity for us to delve deeply into these complex diseases and achieve greater precision in understanding the biology.”
AstraZeneca has multiple cardiovascular drugs and a pipeline of investigational therapies targeting metabolic diseases. However, the field has recently faced setbacks: earlier this year, the therapy named Wainua failed to demonstrate efficacy in a cardiomyopathy trial; last year, the company halted development of a drug in mid-stage trials for the fatty liver disease MASH.
CEO Pascal Soriot is seeking to expand into cardiology and metabolism, with the potential to eventually combine some of these drugs.
Insitro’s differentiation: Before finding the target
“Most AI drug discovery companies focus on what happens after a target has been identified,” said Daphne Koller, CEO of Insitro, in an interview.
Insitro, however, looks further upstream. “What this industry most lacks is a deep understanding of disease biology and knowing where to intervene to truly move the needle,” she said.
She noted that, through this collaboration with AstraZeneca, Insitro’s work on MASH is already quite advanced and is expected to yield drug targets relatively quickly, while cardiovascular research may take longer.
The AI + Pharmaceuticals Boom
This transaction occurred during a week marked by intensive business development activity in AI-driven pharmaceuticals. Just one day before AstraZeneca’s official announcement (October 13), the Danish obesity drug development company Zealand
Pharma announces partnership with generative biology company Generate
In collaboration with Biomedicines, the latter received a $30 million upfront payment, up to $35 million in early preclinical milestone payments, and is eligible for nearly $1.2 billion in total milestone payments; October 14, Vir
Biotechnology has signed an agreement worth up to $1.88 billion with Gilead Sciences; Vir will use its dAIsY AI platform to design drug candidates—just these two deals total over $3 billion in two days.
Looking further ahead, on September 24, Roche's Genentech and Earendil
Labs has entered into a potential partnership with a total value of $1.5 billion; on October 2, Novartis entered into a licensing agreement with Abogen Biosciences for an mRNA pipeline heavily reliant on AI-driven design and sequence optimization, with a potential total transaction value of up to approximately $7.775 billion.
Chinese companies have also stood out in this wave: Since 2026, Insilico Medicine, Hong Kong’s first AI-driven drug discovery company listed on the stock exchange, has secured four partnerships with Servier ($888 million), Eli Lilly ($2.75 billion, including a $115 million upfront payment), SK Biopharmaceuticals ($2.5 billion), and Takeda ($600 million), bringing the total potential value close to $7 billion. Meanwhile, the potential total value of the long-acting peptide collaboration between CStone Pharmaceuticals and AstraZeneca reaches as high as $18.5 billion, with an upfront payment of $1.2 billion. In the first half of the year, Insilico Medicine generated approximately $106 million in revenue, a 287.2% year-over-year increase, and achieved profitability for the first time across all operations.
Tech giants are also moving upstream: NVIDIA and Eli Lilly have established a joint AI innovation lab, and Merck has partnered with Google Cloud on a potential $1 billion Agentic AI deal.
AI collaboration: Novo Nordisk has partnered with both OpenAI and Anthropic. Last month, it was confirmed that Anthropic has built a "wet lab" in the San Francisco Bay Area.
Lab, a physical laboratory for conducting experiments, is extending AI's ambitions into the field of drug science.
These layouts fall on a rapidly heating-up赛道. According to the industry report "AI Drug Discovery Enters the Asset Monetization Era," approximately $60 billion has been invested in AI-driven drug discovery since 2019, involving around 175...
The AI original research project has entered human trials, but no product has yet been approved for market release.
However, AI-driven pharmaceutical companies' revenues currently rely heavily on upfront collaboration payments rather than milestone payments or sales royalties—in this transaction, for example, the $40 million upfront payment represents only about 2% of the total potential value of up to $18.6 billion. The industry's focus is shifting from “model benchmarking” to “clinical validation and drug development capability,” with Moderna and Merck’s personalized mRNA cancer vaccine achieving dual endpoints in its Phase III trial, regarded as the first major validation of AI’s deep involvement in target identification.