1,000+ Opportunities
Find the right grant
Search federal, foundation, and corporate grants with AI — or browse by agency, topic, and state.
The Astra Fellowship is a full-time, in-person research program run by Constellation Institute at its Berkeley research centre, pairing fellows with mentors to work on reducing catastrophic risks from AI.
The January 2026 cohort ran 5 January to 31 March 2026 with an optional extension through the end of June, and applications for that cohort are closed; a subsequent cohort was indicated for Summer 2026, so prospective applicants should watch constellation. org rather than assume an annual cycle. Compensation is 8,400 dollars per month, with empirical research fellows also receiving roughly 15,000 dollars per month in research budget.
Two streams are offered. The empirical research stream covers alignment, control, model evaluations and oversight, and is where the compute budget is concentrated. The strategy and governance stream is aimed at people who already have strong familiarity with AI risk arguments and demonstrated problem-solving ability, and is a genuine entry point for lawyers, economists and policy researchers rather than a token second track.
Prior AI safety experience is explicitly not required for either stream, but relevant technical or domain experience in research, security, governance or policy is. The program targets a specific population - people ready to transition into a full-time AI safety role or to found a new organisation - and its structure reflects that. Over 80 per cent of the first cohort now work full-time in AI safety.
Anyone looking for open-ended research funding rather than a supervised career transition in Berkeley should look elsewhere.
Get a weekly digest of new grants like this
A free weekly digest of new foundation and federal funding opportunities as they're added to Granted. Unsubscribe anytime.
Or search similar grants →According to the current listing, eligibility includes: Open to individuals motivated to address catastrophic risks from AI who have technical or domain-specific experience in research, security, governance or policy. Prior AI safety experience is not required. The program targets people seeking to transition into a full-time AI safety role or to launch a new organisation in the field. Two streams are available: empirical research, covering alignment, control, model evaluations and oversight; and strategy and governance, which expects strong existing familiarity with AI risk arguments and demonstrated problem-solving capability. The fellowship is in person at Constellation's Berkeley research centre and is full time for the duration of the cohort, so relocation is a requirement rather than a preference. The January 2026 cohort ran from 5 January to 31 March 2026 with an optional extension through the end of June, and applications for it are closed. A Summer 2026 cohort was indicated; applicants should verify the opening of the next application window directly at constellation.org, since cohort timing has not followed a fixed annual pattern. Compensation is an 8,400 dollar monthly stipend, with an additional research budget of approximately 15,000 dollars per month for empirical research fellows. Confirm the full requirements in the official notice before applying.
The current listing shows astra pays a stipend of 8,400 US dollars per month. Over the five-month core program that is 42,000 dollars, which is amount_min. Fellows on the empirical research stream additionally receive a research budget of approximately 15,000 dollars per month - roughly 75,000 dollars across five months - so the total value of an empirical fellowship is about 117,000 dollars, which is amount_max. The split matters and is easy to misread: the research budget is compute and experiment money attached to the empirical stream, not additional personal compensation, and strategy and governance fellows should expect the 42,000 dollar figure rather than the headline total. An optional extension through the end of June raises the stipend total further for fellows who continue. The reason to take the program seriously despite it being a fellowship rather than a grant is the placement record: over 80 per cent of Astra's first cohort moved into full-time AI safety roles at organisations including Anthropic, OpenAI and Redwood Research, which is a higher conversion rate than most funded research programs of comparable length. Verify award ceilings, matching requirements, and allowable costs in the official notice.
Constellation Astra Fellowship for Empirical AI Safety Research and AI Strategy and Governance is funded by Constellation Institute. Verify program details on the funder's official page before applying.
Start from the official opportunity page linked in this listing — it carries the sponsor's submission instructions.
Incubator Week is a five-day, all-expenses-paid program in San Francisco run by BlueDot Impact for founders working on catastrophic risks from AI and related technologies, including AI safety, biosecurity and cybersecurity. Cohort 6 runs from 16 to 20 November 2026. On day one, participants build threat models and identify the leading experts in their problem area. Days two and three are for building and testing interventions, consulting experts and stress-testing assumptions with the cohort. Day four is for refining the pitch, and on day five participants pitch for funding. BlueDot funds immediate needs on the spot, and strong progress can unlock a full grant of up to $200,000 within two weeks. No polished idea is required. The Manifund AI Safety Funder Bulletin lists the application deadline as 11 November 2026, though the program page itself does not state a deadline. It suits technically serious people ready to leave their current work to start a new AI safety organization or project.
The K-12 AI Infrastructure Program, run by Digital Promise with Learning Data Insights, DrivenData, Georgetown University's Massive Data Institute and Catalyst @ Penn GSE and funded by the Gates Foundation, released this request for proposals on 7 October 2026. It will fund up to two awards of up to $600,000 each from a $1 million pool. Grantees will build an openly licensed package of tools, workflows and documentation, such as a reference architecture, harness or starter kit, that helps developers explore, test and integrate education-specific digital public goods into AI-enabled formative assessment products. Phase 1 focuses on coordination between teachers and classroom AI tools, giving teachers more visibility into and control over AI. Proposals must cover at least four teacher-AI coordination tasks, engage at least three AI developers and include a classroom demonstration, and most funded work must serve third-party developers rather than end users. Code must be released under MIT or Apache 2.0 and datasets under CC-BY-4.0. The grant period is 18 months in three phases of about six months, starting 1 February 2027. This is a separate, narrower solicitation from the program's March 2026 first grant cycle. Proposals are due 20 November 2026 at 11:59 p.m. Anywhere on Earth.
The K-12 AI Infrastructure Program released this request for proposals on October 7, 2026. It funds an openly licensed package of tools, workflows and documentation, such as a harness, reference architecture or starter kit, that helps developers build education-specific public goods into AI-enabled formative assessment products. The aim is to coordinate the work of teachers and AI tools so teachers can see, control and monitor the AI used in their classrooms. Proposals must name at least four tasks that improve teacher-AI coordination, engage at least three AI developers and include a classroom demonstration application. Most of the effort must serve third-party developers rather than end users. Code must be MIT or Apache 2.0 licensed and datasets CC-BY-4.0. Up to two awards of up to $600,000 each will be made from a $1 million pool. Projects run 18 months in three phases starting February 1, 2027. Proposals are due November 20, 2026 at 11:59 p.m. Anywhere on Earth. This is a separate round from the program's earlier formative assessment, EDU AI and benchmarks RFPs.