GovAI's Entrepreneur-in-Residence is a one-year salaried position on a fixed 12-month term for people who want to identify, test and launch a new AI governance or safety organization, for-profit or non-profit, or an ambitious project within an existing organization. GovAI offers a year of runway and $150k in seed funding via partner funders, plus the support of its team and network, fiscal sponsorship for non-profits and introductions to funders and potential co-founders, and takes no equity or stake in the organizations that emerge. Typical pay is £70,000 to £113,500 a year in London and $80,000 to $180,000 in Washington, DC, and GovAI cannot sponsor visas for the role. Candidates submit an expression of interest, then a pitch of up to two pages, reference checks and one to two interviews. Expressions of interest are reviewed on a rolling basis with no deadline, and the page is dated July 20, 2026.
Programs
Who funds them, what they fund and how to apply.
The Global Technology Risk Foundation provides grants and tailored strategic, operational and network support to individuals and early-stage organizations reducing catastrophic risks from advanced AI. It funds research, compute, career transitions, fellowships and organization building, with no geographic focus or fixed grant size. Applications are reviewed on a rolling basis and require a short proposal and CV.
Kimi's Ambassador Program supports builders, creators and community organizers worldwide who run communities, teach practical AI use and provide product feedback. Benefits include Kimi API credits, resources, community support and collaboration with the Kimi team. Applications have no published closing date or fixed support amount.
Laude Institute runs Slingshots, which it describes as no-strings-attached research collaborations and grants for computer science researchers. Each recipient pledges to ship their research through an open-source project, a startup or another artifact. Support ranges from fast grants and compute to embedded engineering, product support and one-on-one advising; the application form lists a cash grant of $5k to $50k, investment by SAFE, datasets and annotation, AI credits and GPU, and other embedded support. Applicants must be research scientists or engineers in the USA or Canada working on new research. Projects are accepted on a rolling basis with resources deployed immediately, and four batches of recipients have been announced, the latest being Slingshots // FOUR. Applying requires a Laude account.
Leo’s experimental Manifund microgrants offer $10,000 for ambitious projects intended to help AGI go well. Apply only through the specified form, with up to 1,000 words written and typed by the applicant; AI-written or edited applications are rejected. There is no guaranteed response time, and the program can change or stop.
Manifund’s AI Safety Regranting program assigns independent budgets to expert regrantors, who recommend grants using their field knowledge. Manifund reviews recommendations before distributing funds. Projects may include charities, individuals and early-stage work; amounts vary, and regrantor budgets are distinct from individual awards. Projects seek funding by posting a public proposal through Manifund's open call; regrantors choose what to fund.
Manifund's Falcon Fund, led by regrantor Marcus Abramovitch, makes rapid (about one-week), early-stage grants across animal welfare, with a focus on the intersection of animals and transformative AI, such as animal harm benchmarks, animal welfare constitutions for AI systems and watchdog work on AI in farming. Proposals, primarily of $25k to $150k, are submitted as public projects on Manifund on a rolling basis with no deadline; as of September 29, 2026 the fund had $500k to give away and had made 5 grants totalling $140k.
BlueDot Impact funds journalists, writers and creators whose videos, podcasts, journalism and other media help audiences understand risks from advanced AI and biological threats. Media Grants provide $20,000 to $200,000 for 6 to 12 months, covering living and production costs while recipients retain editorial control. Applicants need a track record and an audience; applications have no published deadline.
The NAIRR Pilot is an NSF-led effort, with federal agency and private-sector partners, that gives U.S. researchers and educators access to AI computing, software, data, models and educational resources. Requests are made through nairrpilot.org by researchers and educators at U.S. institutions, including academic institutions, nonprofits, federal and state agencies and startups and small businesses with federal grants. Research Resources requests need a three-page PDF proposal, are reviewed monthly and run for twelve months, and Start-Up Projects offer up to 2,000 GPU-hours on selected systems for three months. Partner contributions listed by NSF include $20 million in Azure compute credits from Microsoft and $30 million from NVIDIA. Proposals are accepted until the end of the pilot or until resources are committed, and on September 1, 2026 NSF announced a NAIRR Operations Center led by SDSC at UC San Diego with TACC at UT Austin.
Coefficient Giving's (formerly Open Philanthropy) fund supporting technical AI safety research, AI governance and policy, and capacity-building to grow the field. Published grants include general support for organizations, compute funds for AI safety research, and individual research projects such as studies on model organisms of reward-hacking, chain-of-thought faithfulness, latent adversarial training, and software formal verification.
Nebius Research Grants gives researchers free GPU cloud credits, Token Factory model tokens and Tavily agentic search credits, along with Nebius expertise. Eligible applicants are postgraduate and PhD students, postdoctoral researchers, faculty leading computational or data-intensive projects at accredited academic institutions, and researchers at accredited academic or non-profit research organizations. Projects need an academic or scientific focus, an innovative use of GPUs for AI, machine learning, data analytics or HPC, demonstrated feasibility, a transparent resource plan and willingness to acknowledge Nebius support publicly. The amount is set individually for each proposal and no fixed figure is published. Applications are open year-round: a complete proposal submitted by the last day of a month is reviewed the next calendar month, and applicants not contacted by the end of that month were not selected. Awards are at Nebius' discretion and credits are issued after the applicant accepts the program terms.
A year-round program offering $5,000 to $50,000 in milestone-based funding, NOS, USDC, and GPU credits for teams building AI applications, infrastructure, and tooling on decentralized compute. Applications are reviewed throughout the year.
NSF CISE's Future Computing Research solicitation supports foundational and interdisciplinary work across algorithms, communications, computing education, systems, human-centered computing, information systems, networking, robust intelligence, software, and hardware. Proposals are accepted at any time, with recurring target dates for the corresponding review cycles.
NSF 26-511 supports US small businesses developing scientific instrumentation and enabling technologies, including AI-driven discovery. Phase I proposals may request up to $305,000 for 6-18 months and require an official Project Pitch invitation. After the July 2026 deadline, the next full-proposal dates are November 4, 2026, March 4, 2027 and July 7, 2027, each at 17:00 in the submitting organisation's local time.
NSF America's Seed Fund supports qualifying US small businesses developing deep technologies. NSF 26-510 provides Phase I grants up to $305,000 and Phase II cooperative agreements up to $1.25 million, with separate Fast-Track and supplemental routes. Phase I full proposals require an official Project Pitch invitation; the next listed deadline is November 4, 2026 at 17:00 in the submitting organisation's local time. Solicitation-wide anticipated funding and award counts are projections, not deployed totals.
Free startup program providing technical resources, training, preferred pricing, partner offers and potential investor introductions. Applications have no fixed deadlines or cohorts. Membership carries no fee or equity requirement; benefits are not a cash award or guaranteed investment.
Sentient Foundation advertises a $42M commitment across two tracks: non-dilutive grants for open-source maintainers, researchers and public-goods builders, and a separate investment track for open-AI companies. Applications are reviewed on a rolling basis. Support can include compute credits, engineering, distribution and community access.
OpenAI's GPT-5.5 Bio Bug Bounty pays vetted bio red-teamers to find a universal jailbreak that defeats OpenAI's predefined biosafety challenge. The original window ran from 23 April 2026 to 22 June 2026, covered GPT-5.5 in Codex Desktop only, and offered awards of $0 to $25,000, with $25,000 for the first universal jailbreak that clears all five questions and smaller awards for partial wins. All prompts, completions, findings and communications are under NDA. An OpenAI page says the bounty is becoming an ongoing private program, the OpenAI Bio Bounty Program, starting with GPT-5.6, with the top reward raised to $50,000, GPT-5.5 testing ending on 27 July 2026, and a rolling application open to past applicants without reapplying.
Status describes whether a program is running. Application methods are recorded separately. Read the funder’s terms before applying.