On October 7, Biohub, The National Institutes of Health (NIH), and U.S. Department of Energy (DOE) announced a major expansion of an international effort to generate open, AI-ready data for predictive models of biology. The partners describe a $1.8 billion combination of funding, data, computation, and new measurement technology. The long-term goal is a virtual cell: a model accurate enough to run experiments digitally and send only the strongest candidates to the lab.
The effort deserves attention from companies that build imaging tools, experimental platforms, data infrastructure, or computational biology capabilities. Most of the investment is expected to go toward producing and organizing the experimental data that predictive models depend on.
What Makes Up the $1.8 Billion
The headline number combines several kinds of contributions, and not all of it is new money:
- Biohub‘s $500 million founding commitment launched the Virtual Biology Initiative in April.
- DOE will invest more than $500 million over five years in laboratory measurement, modeling, and computation.
- NIH will coordinate datasets, repositories, and knowledge bases built through more than $500 million in prior federal investment.
- Google DeepMind, Isomorphic Labs, and Meta are collectively adding $300 million through the Virtual Biology Initiative.
NIH’s role stands out a bit. Datasets built with federal support years ago can gain new scientific value once they are standardized and usable across a larger research community. NIH’s starting point includes the national repositories catalogued by the National Library of Medicine (NLM) and the National Center for Biotechnology Information (NCBI), plus Common Fund programs already building biological atlases and shared data standards.
Measurement Is the Bottleneck
Models can only predict what the data has taught them. Today’s datasets cover a fraction of the cell types and perturbations that matter, so the effort puts its money into generating new measurements. NIH also names a second gap: technologies that can study cells at scales and speeds today’s instruments can’t capture.
Biohub’s investments include cryo-electron tomography, microscopy that can image millions of cells in living tissue, and engineering tools for building and perturbing biology at molecular, cellular, tissue, and whole-organism levels. DOE brings exascale computing, advanced imaging, scattering facilities, and autonomous laboratories through its national laboratory system.
A company with a differentiated imaging method, a reproducible perturbation platform, or an automated experimental workflow may be addressing a constraint the whole effort faces. The technical case needs specifics. What can the platform measure that is hard to observe today? Which cell types and conditions does it cover? How consistent are results across instruments, sites, and runs?
What “AI-Ready” Really Means
The collaboration aims to expand data on how cells respond to interventions across more cell types and conditions. Biohub is working with partners on shared standards, common identifiers, and unified access so datasets can be used together. NIH’s release uses the term SI-ready, short for Super Intelligence, to describe the same goal.
In practice, researchers need to know how the data were generated, which controls were used, how much variation exists, and where the measurements are reliable. Reviewers and partners will probably look at the same things: a defined biological question, sound experimental design, standards that make the output reusable, and independent data showing the results hold up. Rights, consent, and access terms should be settled before anyone builds on the data.
Open Data Comes With Business Decisions
The partners intend to produce an open resource for the research community. Biohub’s head of science, Alex Rives, told Reuters that commercial funders get an embargo period to work with the data before it becomes a public scientific resource, while government-funded work carries no such restriction. Biohub plans to approach pharmaceutical companies and philanthropies next.
Participation decisions start with the business model. A team may be comfortable sharing a benchmark dataset while keeping its instrumentation or software proprietary. Another may need to review customer agreements or data-use restrictions first. These choices affect teaming, proposal scope, and eventual award negotiations, especially when third-party technology or data are essential to the work. An open-data commitment holds up when the plan states what will be shared, when, and how continued use will be supported.
Federal Roles Differ by Agency
DOE’s involvement widens the funding conversation beyond the agencies most life sciences companies already track. Its contribution draws on computing, experimental facilities, modeling, and measurement capabilities. NIH brings biomedical datasets, research programs, and national data infrastructure through its Bio Genesis Mission. The effort also aligns with the Predicting Living Systems National Science and Technology Challenge, which tells you this has national science priority behind it.
A company developing a cell-imaging platform will likely have a different role from one supplying biological models, integrating data, or validating predictions. Some contributions fit a funded research project, others a partnership, and others use of a national laboratory facility.
Rives has said Biohub invites the worldwide scientific community to join the project, and he returns to the main stage at SynBioBeta in May 2027 with an update on progress. EverGlade will be at SynBioBeta, and we’ll be following where outside companies can contribute.
Building the Evidence Behind the Model
The promise of virtual biology depends on predictions holding up when tested. NIH describes the effort as a way to help researchers identify targets and interventions and prioritize promising ideas for laboratory and clinical testing. NIH Deputy Director Nicole Kleinstreuer says universal cell models could bring substantially faster timelines for medical breakthroughs than lab experiments alone.
The most valuable contributions will help researchers measure something reliably, test a prediction, or make evidence usable beyond the team that generated it.
EverGlade helps life sciences and technology companies assess opportunities like this one: where a technology fits, what evidence a program will expect, and what it takes to deliver. Cost planning, data rights, teaming, and post-award requirements belong in that assessment from the start. The detailed opportunities and participation arrangements for this effort are still to come, and we’ll share what we learn along the way. Talk with our team if you want to assess your fit early.
