Roche advances first AI drug discovery project with Recursion Pharmaceuticals

Via advantagecap.com

Roche advances first AI drug discovery project with Recursion Pharmaceuticals

A neuroscience target for an undisclosed neurodegenerative disease marks the first validated output from a partnership worth up to $12 billion

After nearly five years of collaboration, Roche and Recursion Pharmaceuticals have something concrete to show for it. On August 5, 2026, the two companies confirmed that their first AI-identified drug target had been advanced, a neuroscience candidate aimed at treating an undisclosed neurodegenerative disease.

The milestone arrived alongside Recursion’s Q2 2026 financial results. It represents the first validated output from a collaboration that has been running since December 2021, when Roche’s biotech subsidiary Genentech signed on to work with Recursion and handed over $150 million upfront to get started.

Advertisement

What the deal actually looks like

The total potential value is up to $12 billion, spread across as many as 40 programs. Each program carries milestone payments of up to $300 million in development and commercialization fees.

The collaboration covers neuroscience primarily, with oncology included as a secondary focus. Recursion CEO Najat Khan pointed to the company’s AI platform, called Recursion OS, as the engine behind the discovery. The platform works by analyzing large-scale datasets built from cellular imaging and genetic perturbation experiments. For this neuroscience target specifically, the analysis drew on whole-genome CRISPR knockout data from neuronal cells. The platform also incorporates a microglial map built from roughly 46 million images, giving it a detailed picture of the immune cells most relevant to neurodegenerative conditions.

Khan’s argument is that traditional drug discovery relies on biological intuition and incremental hypothesis testing. Recursion’s approach maps biology at a scale that humans cannot manually interpret, then uses machine learning to find patterns in that data that point toward novel targets. The neuroscience result, she said, illustrates that the platform can surface targets that conventional methods have missed.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.

Roche advances first AI drug discovery project with Recursion Pharmaceuticals

Roche advances first AI drug discovery project with Recursion Pharmaceuticals

A neuroscience target for an undisclosed neurodegenerative disease marks the first validated output from a partnership worth up to $12 billion

Via advantagecap.com

After nearly five years of collaboration, Roche and Recursion Pharmaceuticals have something concrete to show for it. On August 5, 2026, the two companies confirmed that their first AI-identified drug target had been advanced, a neuroscience candidate aimed at treating an undisclosed neurodegenerative disease.

The milestone arrived alongside Recursion’s Q2 2026 financial results. It represents the first validated output from a collaboration that has been running since December 2021, when Roche’s biotech subsidiary Genentech signed on to work with Recursion and handed over $150 million upfront to get started.

Advertisement

What the deal actually looks like

The total potential value is up to $12 billion, spread across as many as 40 programs. Each program carries milestone payments of up to $300 million in development and commercialization fees.

The collaboration covers neuroscience primarily, with oncology included as a secondary focus. Recursion CEO Najat Khan pointed to the company’s AI platform, called Recursion OS, as the engine behind the discovery. The platform works by analyzing large-scale datasets built from cellular imaging and genetic perturbation experiments. For this neuroscience target specifically, the analysis drew on whole-genome CRISPR knockout data from neuronal cells. The platform also incorporates a microglial map built from roughly 46 million images, giving it a detailed picture of the immune cells most relevant to neurodegenerative conditions.

Khan’s argument is that traditional drug discovery relies on biological intuition and incremental hypothesis testing. Recursion’s approach maps biology at a scale that humans cannot manually interpret, then uses machine learning to find patterns in that data that point toward novel targets. The neuroscience result, she said, illustrates that the platform can surface targets that conventional methods have missed.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.