Data for AI to understand aging

This round funds projects that create or unlock datasets AI will need to reason about aging and disease. Both new data and more data from existing material count.

New data, for example:

• High-dimensional measurements added to existing trials and longitudinal studies.
• Large-scale perturbation experiments in living organisms.
• Adaptive human trials designed to test AI-generated hypotheses continuously.
• Measurements of how organs and whole organisms change over time in response to interventions.

More data from existing material, for example:

• Tools that extract more biology from samples already collected, such as stored blood or tissue.
• Methods that improve yield from existing biobanks and cohorts, such as the UK Biobank.
• Harmonization or re-measurement that makes existing data usable for training models.

We look for projects where:

• The dataset would not be collected otherwise.
• The data spans interventions, outcomes and/or time courses, so models can learn causal structure.
• The team designs the dataset for AI use from the start, making it open, documented and machine-readable.
• The team can deliver the data quickly.

• Projects that don’t address this round’s data bottleneck. Good science that AI could already do with existing data is out of scope.
• Data collection that depends on humanoid robots or general-purpose lab automation that does not yet exist. We fund data that can be generated using today’s methods.
• Model development, inference tools or analysis platforms without new or newly unlocked data.
• Standard discovery or mechanism studies whose output is a paper rather than a dataset.
• Analyzing an existing dataset with new models, unless the analysis itself produces a new, reusable dataset.

Yes. We fund additional measurements of stored samples, as well as harmonization or re-measurement that makes existing data usable for training models.

Analysis with new models is in scope only if it produces a new, reusable dataset.

We fund projects at any stage, from new ideas to extensions of existing work. The project must produce data that helps AI reason about aging and disease. We are particularly interested in projects that start generating longitudinal data soon.

Established methods: If you are applying well-validated measurement approaches to generate a new or more useful dataset, you don’t need preliminary data. Focus on why the dataset matters and how you will structure it.

New methods or hypotheses: If you propose a new measurement approach or explanation of aging, make the case that it is plausible. Preliminary data is one way to do this, but not the only way. You can also cite supporting literature, pilot results or a strong mechanistic rationale.

We would rather fund the project you are most excited about that could change the field if it works, even if it might fail. We prefer this to work that is certain to produce results but unlikely to move the field forward.

Researchers work within a publishing system that rewards positive results, which can make ambitious projects feel risky. To help funded projects share all their findings, we have previously worked with GeroScience to enable short reports from Impetus-funded projects with null results.

Our grants range from $10K to $500K. Reviewers consider the amount you request: if two projects are otherwise equal, they will favor the one that needs less funding. On top of your requested amount, we pay institutional overhead of up to 10%.

Yes. People are the most essential part of any research project, and usually the most expensive. Feel free to use grant funds for salaries.

Using a large fraction of the grant for equipment purchases is not in scope. Plans that depend on tools that have yet to be invented are out of scope.

Academic and nonprofit researchers around the world can apply, at any career stage, including trainees and students. Unaffiliated researchers are also eligible. All applicants must have access to the facilities and resources needed for the proposed work.

You don’t need to work in aging biology already. We welcome applicants from other fields whose projects will produce data relevant to aging, disease or treatment outcomes.

No, we do not fund for-profit efforts.

We review applications in three stages so we can handle a large volume of proposals.

1. Short proposal. Submit about three paragraphs describing your idea. Longevity experts check that each proposal fits the round’s goals and could be impactful if successful.
2. Detailed plan. Shortlisted applicants submit 1–2 pages describing how they will carry out the work and their goals and plans for open-source datasets. A biologist and a machine learning researcher review each proposal. Sometimes an expert in the proposed method also reviews it. They assess technical feasibility, whether the scientific plan is sound for the intended outcomes, and whether the proposed data structure suits use in AI models.
3. Interview. Applicants we don’t already know have a fifteen-minute chat with Impetus Grants founder Martin Borch Jensen.

You can access the Stage 1 application [application link].

For Round 4, we aim to give every applicant a final funding decision within five weeks of receiving their application.

The three-week figure elsewhere on this page describes Impetus’s historical decision timeline.

We don’t set a fixed project period. Propose whatever timeline your project realistically needs, and explain it in your application if necessary. We favor proposals that deliver results sooner.

Yes, you can submit multiple proposals. Each applicant can receive up to $500K in total across Round 4 awards, including awards as a co-applicant.

Yes, you can include co-applicants. If we fund your project, we pay the grant to the lead applicant’s institution, which can subgrant part of it to a co-applicant’s institution.

Please disclose both active and pending grants in the “other funding” section.