The race to turn AI drug discovery from laboratory concept into actual medicines is heating up, with major pharmaceutical players and tech companies making bold moves to industrialize the field. Japanese biotech firm Nxera Pharma has joined OpenFold, a non-profit AI research consortium backed by tech giants like Amazon Web Services, Microsoft, and NVIDIA, while ByteDance's AI pharmaceutical unit is spinning off as an independent company to pursue financing and accelerate its path to market.

These parallel developments signal a critical inflection point: AI drug discovery is transitioning from academic research into the messy, complex world of actual pharmaceutical development. The moves reflect growing confidence that AI can help solve one of the industry's most stubborn problems: the astronomical cost and time required to bring new medicines to patients.

Why Are Pharma Companies Joining AI Consortiums Now?

Nxera's decision to join OpenFold represents a strategic bet that open-source AI tools will accelerate its internal drug discovery efforts. The company, which specializes in a challenging drug target class called GPCRs (G-protein coupled receptors), sees the consortium as a way to access cutting-edge structural prediction models while contributing its own computational chemistry expertise. Nxera is participating as a user of OpenFold's tools and is not required to share proprietary data or intellectual property, a key incentive for pharmaceutical companies historically protective of their research.

"Our participation in the OpenFold Consortium is highly aligned with the core design philosophy of the NxWave platform: the integration of structural biology, computational chemistry and AI. Access to an ecosystem that advances open-source AI drug discovery alongside world-class pharmaceutical companies and research institutions will further strengthen our AI-enabled drug discovery capabilities for GPCRs, a medically important and challenging target class," stated Patrik Foerch, Chief Scientific Officer and President of Nxera Pharma UK.

Patrik Foerch, Chief Scientific Officer and President of Nxera Pharma UK

OpenFold itself has grown into a sprawling international ecosystem that includes major pharmaceutical names like Bristol Myers Squibb, Novo Nordisk, Bayer, and Roche, alongside academic institutions and biotech firms. The consortium's mission is straightforward: provide researchers with access to powerful AI-enabled tools to create medicines that would be impossible to discover without AI.

What Is ByteDance's AI Pharmaceutical Unit Doing Differently?

ByteDance's approach is more aggressive. The Chinese tech giant's AI pharmaceutical team, established in 2021 and led by Liu Kai, is spinning off as an independent company to pursue dedicated financing and establish what the company calls an "independent organizational structure that better suits the characteristics of this business". The core team of about 50 people, composed of AI algorithm specialists and senior pharmaceutical experts, will transfer to the new entity along with the company's technology platform and existing drug pipeline assets.

ByteDance's AI pharmaceutical capabilities have advanced significantly beyond basic research. In 2025, the team released molecular structure prediction models called Protenix and Seedfold, which it iterated to Protenix-v1 and v2 in 2026, building high-precision open-source structure prediction capabilities for biological systems like proteins and ligands. The company has also launched protein design tools and an AI pharmaceutical platform called Anew Labs for real-world drug research and development.

The most concrete evidence of progress is ByteDance's early-stage drug pipeline. In April 2026, Anew Labs disclosed an IL-17 small-molecule project that achieved something significant: the blockade of three dimers (AA/AF/FF) of the IL-17 family with small molecules for the first time globally. IL-17 is an important pathway for autoimmune diseases like psoriasis and ankylosing spondylitis, making this a potentially valuable therapeutic target.

How Are Companies Positioning AI Drug Discovery for Industrial Scale?

• Open-Source Collaboration: Companies like Nxera are accessing state-of-the-art structural prediction models including OpenFold and OpenFold3-preview through consortium participation, allowing them to evaluate and benchmark these tools against their own proprietary approaches without sharing sensitive intellectual property.
• Organizational Independence: ByteDance is establishing a separate legal entity with its own governance structure to attract top-tier pharmaceutical talent and create decision-making flexibility that differs from typical internet business operations, recognizing that drug development requires different management approaches.
• Computing Infrastructure Support: ByteDance's new pharmaceutical company will continue receiving computing power support from Volcengine, the company's cloud computing subsidiary, ensuring the AI models have the computational resources needed for complex molecular simulations and predictions.

The pharmaceutical industry faces enormous pressure to improve efficiency. Global drug expenditure is projected to reach approximately $2.3 trillion by 2028, according to IQVIA, one of the world's largest medical health data and clinical research service companies. Yet the core pain points of high research and development costs, long development cycles, and high failure rates for new drugs have not fundamentally changed. The industry is eager to introduce AI technology to break through these limitations.

The progress in AI models themselves demonstrates why pharmaceutical companies are investing now. Google DeepMind's AlphaFold series shows the trajectory: from initial feasibility verification to AlphaFold 2 achieving atomic-level precision prediction of 200 million proteins, and then to AlphaFold 3 breaking through the limitation of single proteins to accurately predict complex interaction systems. If protein structure prediction represents basic research, then the multi-modal molecular generation models emerging in recent years directly address the core problem of the pharmaceutical industry: drug design.

Both Nxera and ByteDance are betting that AI has now penetrated deep enough into drug discovery that the field is ready to move from research into industrial application. The next phase will test whether these companies can translate algorithmic breakthroughs into medicines that actually reach patients and improve human health.