AstraZeneca Acquires Modella AI to Integrate Foundation Models into Global Oncology R&D

AstraZeneca has acquired Boston-based Modella AI to scale the integration of multimodal foundation models and AI agents across its oncology research and development programs. The deal follows a collaboration between the two companies initiated in July 2025, and brings Modella’s AI technologies directly into AstraZeneca’s clinical and preclinical infrastructure. No financial terms were disclosed. 

Modella AI specializes in generative and agentic AI platforms trained on clinical, pathological, and molecular datasets. The company has focused on oncology applications such as biomarker discovery, digital pathology, and predictive modeling. Its acquisition is intended to accelerate trial design, data interpretation, and global deployment of AI-driven decision support in AstraZeneca's oncology pipeline.

Modella’s technology stack includes pathology models and agent-based systems:

• PathChat is a digital assistant for pathologists and researchers that helps interpret pathology slides. It can look at both images and text at the same time, making it easier to describe what’s seen under the microscope, answer open-ended questions, and draft reports. Users can upload multiple high-resolution images and interact with the model directly inside a slide viewer. The latest version, PathChat 2, has improved abilities in identifying disease patterns, describing tissue features, and following instructions in research and training settings.
• Judith is an AI tool that automates the process of building image analysis models for medical research. Instead of coding or manually adjusting models, users just describe what they want to do like identifying certain cells or structures in tissue images. Judith takes care of setting up the model, running the analysis, and interpreting results. It supports large and complex images, such as whole-slide scans, and can also help uncover biomarkers using pretrained foundation models.

Modella AI joins a broader set of AI partnerships and platforms that AstraZeneca is using to feed this strategy. In cancer, the company is already co-developing a large multimodal foundation model with Tempus and Pathos AI, and recently signed a separate collaboration with BostonGene to apply its tumor–immune foundation model for patient stratification and trial design. 

Beyond oncology, long-running target discovery work with BenevolentAI in kidney, lung, autoimmune, and heart diseases, as well as AI-enabled gene therapy discovery with Algen Biotechnologies, small-molecule discovery with CSPC Pharmaceuticals, and immune target discovery in IBD with Immunai, points to a consistent pattern of using external AI platforms alongside in-house tools across multiple disease areas. For further context on how Big Pharma is deploying AI across the R&D pipeline, see our deep dive Inside Big Pharma's AI Playbook: From Molecule Discovery to Clinical Trials.

Topic: Industry Movers