CFAs: AI for Science and Safety Nodes RFP Local Compute

Deadline: 31-Oct-2026

Foresight Institute is inviting proposals for projects that build and scale secure, locally owned AI compute clusters to strengthen open science and AI safety research. Grants typically range from $30,000 to $100,000 and can support short-term or multi-year projects focused on deployable AI infrastructure, accessible node-management software, sustainable hardware models, and scalable approaches that enable individuals and communities to own and control their computing resources. Foresight Institute has a longstanding focus on openness in science and technology and supports the development of independent ecosystems for scientific research and AI safety. The RFP responds to concerns about the concentration of AI compute resources among a relatively small number of companies. The program seeks projects that can broaden access to computing infrastructure and enable researchers and communities to develop greater control over their own AI compute resources. The focus is on building practical, secure, and scalable local compute setups that can support AI research while contributing to open science and AI safety.

Selected projects can typically receive:

The funding structure may depend on the project design, milestones, and grant arrangements. The RFP seeks innovative solutions for building and scaling local AI compute infrastructure.

Potential project areas include:

Projects should focus on practical approaches that can expand access to independent AI computing infrastructure. The program is based on the idea that access to AI computing resources is an important part of maintaining an open and independent research ecosystem. When advanced compute resources are concentrated among a limited number of organizations, researchers and communities may face challenges related to: Local AI compute clusters could provide an alternative by allowing communities and organizations to develop and manage their own infrastructure. The RFP welcomes both technical and economic approaches to local AI compute.

Examples include:

These examples illustrate the types of challenges the RFP is seeking to address. Both nonprofit and for-profit entities are eligible to apply. Eligible applicants may include organizations or teams capable of developing solutions aligned with the RFP’s focus on local AI compute infrastructure. For-profit applicants must be able to justify why grant funding is needed for their proposed project.

Proposals should demonstrate:

The program encourages high-risk, high-reward ideas, particularly projects that could achieve significant progress within relatively short timelines. All funded work products must be open-sourced. Applicants should therefore consider open-source development as a core part of their project design rather than treating it as an optional feature. Proposals should explain how project outputs can be shared openly and contribute to the broader AI research and safety ecosystem. Grant payments may be provided in different ways depending on the project.

Possible arrangements include:

For multi-year projects, funding may be structured around project milestones and progress. Up to 10% of direct research costs can be allocated to overhead. Successful applicants must complete due diligence checks before receiving funding. Grantees are also expected to meet basic reporting requirements, including: Applicants should be prepared to provide the information required for due diligence and ongoing grant administration. The RFP encourages ambitious ideas that may involve significant technical or implementation risks but could produce substantial benefits if successful. Projects should balance ambition with feasibility and demonstrate why the proposed approach can produce meaningful results within short AGI timelines. The program is particularly interested in solutions that could create durable, scalable infrastructure for independent AI research and safety. Applicants should build their proposal around the specific local compute challenge they intend to address.

Important preparation areas include:

Applicants should avoid:

Applicants can strengthen their proposals by: It is a Foresight Institute funding opportunity for projects that build and scale secure, locally owned AI compute clusters to support open science and AI safety research. Grants typically range from $30,000 to $100,000. Yes. Funding is available for both short-term and multi-year projects. Yes. Payments may be made as a lump sum or in tranches based on project milestone completion. Up to 10% of direct research costs may be allocated to overhead. Yes. For-profit entities can apply, but they must justify their need for grant funding. The RFP encourages projects such as deployable AI cluster prototypes, node-management software, sustainable hardware models, and scalable approaches for distributing AI clusters across different environments. Yes. All funded work products are required to be open-sourced. Grantees must provide regular progress updates and publicly acknowledge their project unless privacy is requested. Yes. The RFP explicitly encourages high-risk, high-reward ideas that could have significant impact and are feasible within short AGI timelines. The program aims to strengthen independent AI research ecosystems by enabling individuals and communities to own and control their own compute infrastructure rather than relying entirely on concentrated external compute resources. The Local AI Compute Clusters RFP provides typically $30,000–$100,000 in funding for projects developing secure, scalable, and locally controlled AI computing infrastructure. By supporting open-source technical and economic solutions, the program seeks to expand independent access to AI compute and strengthen open science and AI safety research. For more information, visit Foresight Institute.