Drug collaborations don’t always work out as planned. Sometimes they work out better.

When Incyte agreed last year to partner with artificial intelligence (AI) platform developer Genesis Molecular AI to research, discover, and develop at least two small molecule treatments, they designed a collaboration that would generate at least up to $620 million for Genesis, whose foundation models for molecular AI are designed to power agentic drug design and development.

The companies now say they made enough progress over the past 15 months to expand their AI-based drug collaboration to encompass at least five targets—with a potential payoff for Genesis that has ballooned to over $1 billion.

Behind that expansion, Incyte and Genesis say, is the promise shown so far by the two initial targets, both selected by Incyte as called for in the initial strategic collaboration. One is a “very hard-to-drug, novel target” for which the companies worked to create novel, first-in-class chemical matter, while the other is a target that other companies have sought to make druggable without success, Pablo J. Cagnoni, MD, Incyte’s president and global head of R&D, told GEN.

“Novel targets create problems for obvious reasons. You don’t have any chemical matter that you know to start with. The collaboration with Genesis has jump-started that program significantly,” Cagnoni said of the first target. “You need a crystal structure, you need to know which particular site in the target you need to bind, and then you need to start making chemical substance against it.”

“It’s easy to make chemical matter, it’s really hard to make medicines—so that was the optimization step that Genesis really helped us do,” Cagnoni added.

The second target, he explained, required not only high potency and very high selectivity, but unique pharmaceutical and pharmacokinetic properties. The companies were able to incorporate those and other properties for the target with help from Genesis’s generative and predictive AI platform, Genesis Exploration of Molecular Space (GEMS).

GEMS integrates AI and physics into models designed to generate and optimize drug molecules. GEMS’ generative diffusion model for structure prediction, Pearl—short for “Placing Every Atom in the Right Location”—was unveiled in an October 26 preprint showing it to have surpassed AlphaFold 3 and other open source baseline models on the public protein-ligand co-folding benchmark Runs N’ Poses (14.5% improvement) and the docking and molecular generation benchmark PoseBusters (14.2% improvement).

‘Substantial progress’

“By being able to optimize multiple parameters at the same time with the help of the GEMS platform and our colleagues at Genesis, we were able to really make substantial progress that was eluding us with other technology,” Cagnoni said. “The collaboration with Genesis has allowed us to make significant progress on the path to an IND. We’re not quite there, but we’re getting pretty close to that.”

The two targets, he said, represented opposite ends of the drug discovery spectrum: “For one, we had something that started to look like a drug but wasn’t good enough. For the other one, we had a great target and no drugs. So, taking a view of those two ends of the spectrum, convinced me that we had to expand this, make it as broad as possible, and that’s why we put in place a new collaboration.”

As with their initial collaboration, the companies aren’t yet revealing the targets or therapeutic areas in which they are working, though Cagnoni said they fall within one of Incyte’s three current therapeutic areas of interest: hematology, oncology, and inflammation and autoimmunity, a narrower niche within the traditional I&I (inflammation and immunology) focus area.

Through the expanded collaboration, Incyte will use its proprietary experimental data to train Genesis’ GEMS platform, with the aim of accelerating drug development across multiple programs.

Options beyond five targets

Incyte will select at least five new targets to develop with Genesis, with options to nominate additional collaboration targets over time. Incyte will have exclusive rights to develop and commercialize treatments developed through the collaboration.

“We know what properties a priori we need to optimize for, always with some caveats,” Feinberg said. “We almost always know that we need to achieve potency, selectivity, a wide variety of ADME [absorption, distribution, metabolism, and excretion] properties. Usually, in a given program, something like 30 or so different ADME assays are routinely run to some degree of frequency. This can often feel like playing whack-a-mole, instead of the serious engineering task of multi-parameter optimization.”

“Our aim,” he added, “is to render the drug discovery process as much like the latter and as little like the former.”

Incyte has agreed to pay Genesis $120 million upfront—to consist of $80 million cash and a $40 million purchase of Genesis’ equity—and unspecified recurring research funding to support AI model training and inference computing. Incyte has also agreed to pay genesis up to $232 million in payments per target, tied to achieving preclinical and clinical development, regulatory, and sales milestones.

The collaboration is the second AI-focused partnership announced by Incyte in late May. A day before the Genesis expansion announcement, Incyte said it had launched a separate strategic collaboration with Edison Scientific to employ its Kosmos AI platform for discovery and development work—namely enabling continuous learning from translational and clinical data, real-time synthesis of evidence and predictive models of therapeutic performance.

Incyte and Edison disclosed the focus of their initial project: “high-impact” use cases in target discovery and validation and translational biology, where Edison’s AI capabilities will be embedded within Incyte’s research workflows. The companies said they aim to support more efficient exploration of experimental, clinical, and biomarker data with the potential to expand across Incyte’s broader R&D organization.

As for Incyte’s collaboration with Genesis, if Genesis achieves all milestones across the five initial targets of the expanded partnership, including multiple indications and major territories, Incyte will pay the company more than $1 billion—as long as the aggregate peak annual net sales of the five products exceed specified milestones. Payments could grow to “several” billion dollars depending on how many additional collaboration targets are nominated, and how many milestones are achieved.

Genesis is also eligible to receive royalties on sales of any approved collaboration products.

Stanford spinout

Genesis spun out in 2019 from the Stanford University lab of Vijay Pande, PhD, co-founder and managing partner of the venture capital firm VZVC and a former general partner at Andreessen Horowitz (a16z) and founding general partner of its bio funds. Feinberg was a graduate student in Pande’s lab who co-invented and co-authored key peer-reviewed papers detailing deep learning technologies.

In 2020, Genesis won a $52 million Series A financing. The company has grown since then to raise $340 million, most of that consisting of $200 million Series B financing completed three years later, plus the $40 million strategic investment Incyte made in Genesis equity as part of the companies’ expanded partnership.

In addition to a16z, Genesis’ investors have included NVentures, the venture capital arm of AI chip giant Nvidia, which has expanded in recent years into biopharma among other industries.

Incyte is the fourth and latest biopharma giant to partner with Genesis on an AI-focused drug discovery and development collaboration applying GEMS. Genesis garnered $35 million upfront in launching its partnership with Gilead Sciences in 2024, and earlier announced past collaborations with Eli Lilly and Genentech, a Member of the Roche Group.

“Our mission at Genesis is to create AI technologies that enable creating drugs that otherwise would not be possible,” Evan Feinberg, PhD, Genesis’ founder and CEO, told GEN. “And thanks to working with really, really elite drug discovery teams, like what Incyte has, we’re able to work on a wide spectrum of very important problems in drug discovery.”

That work, he asserted, requires discerning the uniqueness of each potential target.

“Every target is really its own special snowflake in some way. Every drug target really entails its own challenges, oftentimes requires its own special approach,” Feinberg said. “Over the past year, we were able to work on two very different programs, that each have their own challenges, and thereby enable us to adapt and deploy our GEMS AI platform in these very different settings, bringing one of those two targets much closer to IND, and for the other target finding the first-in-class chemical matter, which was a very exciting year of work.

“Now we’re excited to address the challenges ahead with this, expanded partnership together,” Feniberg added.

The post Small Molecules to Big Partnership: Incyte, Genesis Expand AI Collaboration to $1B+ appeared first on GEN – Genetic Engineering and Biotechnology News.