Turbine Raises $25M to Scale AI Digital Cell Models Into Immunology, Following AstraZeneca Collaboration
Turbine, London- and Budapest-based virtual biology company building AI-driven “digital cell” models, has closed a $25 million Series B round. It was led by Interactive Venture Partners, with participation from Beiersdorf Venture Capital and existing investors including MSD Global Health Innovation, Accel and Mercia.
Alongside the financing, the company disclosed its first immunology-focused collaboration with an unnamed top 10 pharmaceutical company, extending its virtual cell platform beyond oncology into immune-mediated disease research. Previously, in 2025, Turbine announced an antibody-drug conjugate (ADC) discovery collaboration with AstraZeneca.
The newly announced immunology collaboration will focus on modeling immune cell behavior using proprietary datasets supplied by the pharma partner. Turbine’s platform will be trained to represent immune pathways and to simulate potential drug combinations. The stated objective is to prioritize combinations and patient subsets for further validation, addressing the combinatorial complexity typical of immunological diseases where multiple pathways intersect. The exact collaboration partner has not yet been disclosed.
The new capital is intended to expand Turbine’s virtual cell platform across additional assay types in discovery and translational medicine. The company operates what it describes as a lab-in-the-loop model, where proprietary perturbation datasets generated in wet lab settings are used to iteratively refine its foundational virtual cell model. These models are used to create computational replicas of biological assays, which can then be run at scale through Turbine’s Virtual Lab, a no-code interface designed to integrate with pharmaceutical R&D workflows.
The concept of a “virtual cell” dates back to early computational biology efforts in the late 1990s and early 2000s, when platforms such as VCell and CompuCell3D used mechanistic equations and stochastic simulations to model selected cellular processes.
A major milestone came in 2012, when Markus Covert’s group at Stanford published the first whole-cell model of Mycoplasma genitalium, integrating all 525 known genes of the organism into a single computational framework. While foundational, these mechanistic approaches proved difficult to scale to more complex human cells due to multi-scale interactions, nonlinear dynamics, and limited standardized datasets.
A new phase began with the integration of large-scale single-cell datasets and machine learning. In late 2024, a Cell study titled “How to build the virtual cell with artificial intelligence: Priorities and opportunities,” authored by an interdisciplinary team including researchers from EPFL, Stanford, Genentech, Google Research, Harvard, and the Chan Zuckerberg Initiative (CZI), formalized the concept of the AI Virtual Cell (AIVC).
The framework proposes interconnected foundation models that learn generalizable biological representations across modalities and scales. CZI has been a central force in this shift, developing large-scale models and launching the Billion Cells Project to generate training data at unprecedented scale. Together, these efforts mark a transition from handcrafted mechanistic simulations toward data-driven, multimodal models designed to enable in silico experimentation and, ultimately, patient-specific virtual twins.
For a broader analysis of how foundation models, perturbation atlases, and lab-in-the-loop systems are reshaping the virtual cell landscape, see our recent deep dive, Building the Virtual Cell: AI Foundation Models and Billion-Cell Datasets
Turbine’s approach centers on simulating how cells respond to genetic or pharmacological perturbations, with the goal of narrowing experimental space before committing to physical validation.
Image credit: Turbine
According to the company, its virtual assays have been applied across more than 30 discovery programs at biopharma partners including MSD, AstraZeneca and Bayer.
Topic: Biotech Ventures
