The frontier should reach every lab.
The future of intelligence should be shaped by the people who live in it. We want researchers, builders, and communities to have the means to develop AI, understand its behavior, and have a voice in how it is used. Open research was what helped bring us here, from the Transformer paper to the models and enormously valuable enterprises built upon it. Yet today the compute, funding, talent, and research needed to do the same are concentrated within a few institutions, and the largest labs publish a fraction of what they learn.
We want the tradition of open research to continue to inspire generations of great researchers and pioneers. That ethos led us to create Democratize Intelligence (DEMi), our conference series and network dedicated to open intelligence. It brings together people working on new computing architectures, open models, alignment, and public participation. Our mission to expand access to compute grew from helping people there find the resources and collaborators they needed.
Epilogue supports open research by academics, independent researchers, and small labs through compute grants. The results are shared publicly, so others can extend them and the frontier stays open to everyone who wants to explore it.
About Democratize Intelligence →Independent research grants
The pool supports open research experiments, including work in academia. We also offer discounted compute for early teams doing interesting work in the space.
- 01Research grantsGrants provide free or discounted compute for selected open, non-commercial research by academics, independent researchers, and small labs. A proposal should describe the research question, the experiments an allocation would support, and the work that will be made public. That may include model weights, code, or published findings that other researchers can evaluate and use.
- 02Compute for early teamsWe offer discounted blocks to exceptional teams at a very early stage. Share the workload, hardware requirements, and expected duration so we can discuss a suitable allocation and its terms.
- 03The shared poolEpilogue contracts for the underlying capacity and manages allocations across projects with different schedules and requirements. When a project’s needs change, its unused months can return to the market.
Building with Lightyear.
Lightyear is a residency in SF incubated by HF0 for exceptional founders on a hero’s journey: the Olympic Village for founders. The accelerator is designed to make thirty days in San Francisco the most productive month of their lives. We are the official compute partner for Lightyear, serving teams doing frontier work in robotics, physical intelligence, and world models.
If you are a current or former Lightyear team, you are welcome to join our compute pool. Use the form below to discuss your needs.
About the residency →Apply for a grant or request capacity.
Describe the research or workload, the compute it requires, and the proposed timeline. For a grant, explain what you intend to publish openly and confirm the work is non-commercial. Include a paper, repository, or other relevant work where available. We will follow up by email.

