Takeda Expands AI-Driven Antibody Design Partnership with Nabla Bio in a Billion-Dollar Deal
U.S. biotech company Nabla Bio has signed a second major research partnership with Japan’s Takeda Pharmaceutical Co. (4502.T), expanding their collaboration in artificial intelligence-driven drug discovery. The new multi-year agreement builds on an initial collaboration launched in 2022 and could be worth more than $1 billion in success-based payments, alongside upfront and research funding in the tens of millions.
At the core of the deal is Nabla’s proprietary AI platform, the Joint Atomic Model (JAM), which designs protein-based therapeutics de novo. Unlike traditional antibody discovery methods that rely on screening natural repertoires, JAM generates entirely new antibody sequences optimized for affinity, manufacturability, and drug-like properties. According to Nabla, its integrated wet-lab validation enables a design-to-experiment feedback loop of just three to four weeks, which the company claims is the fastest in the industry.
Technical results disclosed alongside the announcement suggest significant progress in overcoming long-standing challenges in de novo antibody design. Nabla reports double-digit hit rates, picomolar-level binding affinities, favorable pharmacokinetics (pK), and low immunogenicity in both animal models and non-human primates (NHP). If validated, these results would mark the first demonstration of such performance for AI-designed antibodies and suggest that first-in-human trials may begin within one to two years.
The expanded partnership will focus on hard-to-treat diseases, including the development of multi-specific antibodies, receptor decoys, and other custom biologics. Nabla CEO Surge Biswas emphasized that the collaboration targets the most pressing scientific barriers within Takeda’s early discovery pipeline, leveraging JAM to accelerate therapeutic development.
The deal comes amid broader industry momentum around AI in biopharmaceutical R&D. Earlier this month, Takeda joined a consortium with Bristol Myers Squibb and others to train large-scale AI models on shared datasets, while also announcing a shift away from cell therapy research toward more scalable therapeutic modalities.
Also, AstraZeneca has entered a collaboration worth up to $555 million in regulatory and commercial milestones with San Francisco–based Algen Biotechnologies, gaining exclusive rights to develop and commercialize therapies using CRISPR gene-editing technology.
In the meantime, Sanofi has entered a three-year licensing deal with Toronto-based BenchSci to roll out the company’s ASCEND platform across its global preclinical research teams. The agreement integrates neurosymbolic AI into Sanofi’s core R&D operations, giving its scientists and contractors access to ASCEND’s structured discovery environment.The
BiopharmaTrend team explored this topic in a recent newsletter issue, Inside Big Pharma's AI Playbook: From Molecule Discovery to Clinical Trials, which you can read on Substack.
If successful, the Nabla–Takeda program could help close the translational gap between AI-generated protein designs and safe, effective human therapeutics, an inflection point many in the field have long anticipated...
Topic: Industry Movers
