Four Pharma Rivals Built an AI Sandbox Together—Then Moved to Quantum

In December 2020, four of the world's largest pharmaceutical companies did something that should have been impossible: they agreed to build a company together. Pfizer, AstraZeneca, Merck, and Teva are direct competitors. They hunt the same diseases, chase the same market opportunities, and fight for the same patients. Getting two of them to share equipment would be unusual. Getting all four to co-found and jointly own the same venture studio should never happen.

Yet it did, in Israel, and it worked. The reason was timing. In 2020, artificial intelligence was clearly about to reshape drug discovery, but nobody inside pharma knew which approaches would actually work. The tools were expensive. The talent was scarce. A wrong bet could burn through hundreds of millions in R&D spending. Four competitors found a way to explore the unknown together without exposing their proprietary work. They won a competitive tender from the Israel Innovation Authority and built AION Labs—a venture studio funded and operated jointly, with AWS as technology partner and the Israel Biotech Fund as venture backer. The structure was elegant: rivals could never pool their actual drug development, but they could share the cost and the learning of testing whether a given AI idea had merit, as long as it happened one step removed, inside independent startups on neutral ground.

Israel was the natural choice. The country had deep reserves of AI engineering talent, and neither AstraZeneca nor Pfizer had ever run R&D operations there—only commercial ones—so no company had turf to defend. It was neutral ground for rivals who competed everywhere else. What they built there operates less like a traditional incubator and more like a matchmaker. When one of the four pharma partners surfaced a hard problem, AION would find a scientist to build a company around it, often recruiting straight from academia. The studio would hand over roughly a million dollars, two years of runway, a borrowed CFO, and direct access to pharma R&D that a normal startup would spend years trying to open. In return, the four partners fed in challenges, sat on investment committees, and co-developed from day one. Because putting four competitors around one table meant no one could build against a shared drug target, AION built platforms instead of single assets. That constraint, born of antitrust law, ended up defining the entire portfolio.

Line up the startups by launch date and the portfolio reads like a timeline of AI hype. In 2022, OMEC.AI arrived with the most 2022 idea imaginable: use machine learning to predict which drugs would survive clinical trials by analyzing messy preclinical data. Its founders came straight from Mobileye, Intel's self-driving unit—the purest expression of the moment's belief that AI talent from any domain could be transplanted into biology. A year later, generative protein design was the wave, riding directly behind AlphaFold's breakthrough. DenovAI launched to design antibodies entirely from scratch, licensing technology from EMBL. By 2024, targeted protein degradation was fashionable, and TenAces appeared to hunt molecular glues with machine learning, promising to reach the roughly 85 percent of proteins considered undruggable. Later that year came ProPhet, speaking fluently in the language of diffusion models and large language models, aimed at generating small molecules for targets with almost no structural data. Around them sat companies focused on gene editing, tumor heterogeneity, and epiproteomics—each one a small time capsule of whatever pharma was chasing the month it launched.

The experiment paid off in January 2026. CombinAble.AI, an antibody-optimization startup founded inside AION in 2023 by Daria Kokh and Noam Katz, was acquired by insitro, Daphne Koller's San Francisco AI-therapeutics company, which opened a new Israel R&D center around the team. CombinAble was built from scratch with no background intellectual property and no existing team, and it reached acquisition in a little over two years. For a studio whose entire premise is that four rivals can manufacture companies from nothing, that was the clearest proof the machine works. Across the portfolio, AION points to more than 22 proofs-of-concept run with its partners and over 30 million dollars in total funding. DenovAI and Promise Bio are both raising Series A rounds, early signs that CombinAble's exit is not a one-off but the leading edge of a cohort reaching that stage. Two of the nine companies have closed. Not everything works, and the studio's leadership is blunt about the recurring flaw: most of these companies struggle to find a business model that isn't simply making a drug.

But here is the paradox: AION Labs could not be built today. Every major pharmaceutical company now runs its own AI unit. The shared blind spot that pulled four rivals into one sandbox in 2020 has closed, precisely because sandboxes like this one helped close it. The experiment worked well enough to put itself out of a job. So AION went looking for the next thing pharma is too uncertain to touch alone. This time it is quantum computing, and the setup rhymes with 2020: loud with promise, no clear use case, no one inside pharma sure what to do with it. Some people say it is bullshit, and in just twenty years we will see something. Nobody, including the companies funding the effort, knows whether quantum will matter in five years or twenty. That uncertainty is the qualification, not the disqualification. It is exactly the kind of bet the sandbox exists to make cheaply. In 2020 the shared bet was AI, and it paid off with an exit on the board. In 2026 it is quantum, same structure, same logic, same four companies splitting the risk.