Currently 6 active ai and ai safety grants, fellowships and residencies open to applicants in the US. Hand-curated and updated weekly. Every entry is funded, no exposure-only calls. Browse the list below, or use the interactive desk for filtering and shortlisting.
Subscribe: RSS feed · Calendar (.ics)
Fully funded in-person fellowship pairing fellows with senior mentors to do machine-learning research and engineering on core technical AI safety topics: AI alignment, AI control, model evaluations, scalable oversight. Mentors are drawn from partner organisations (full list on the Astra site). Includes weekly mentor and research-manager meetings, talent mobilisation/career support, startup incubation services, and access to a world-leading AI safety convening space (300+ weekly visits). Visa support: J-1; F-1/CPT (full-time only) or OPT, with DSO approval. Looking for: strong Python proficiency, agency and proactivity, basic ML and AI safety knowledge, ability to learn and iterate quickly. Application: resume + at least two references (Constellation network preferred) + optional CV/website/written samples. Stages: initial screening, coding test + reference check, work tests, interview. Deadline 3 May 2026, 23:59 Anywhere on Earth. Contact astra@constellation.org. Over 80% of the first cohort moved into full-time AI safety roles at OpenAI, Anthropic, Google DeepMind, Redwood Research, METR, CAISI and the UK AI Security Institute.
Fully funded in-person fellowship to develop and pursue research projects that reduce catastrophic risks from AI, on the strategy and governance side rather than the technical/empirical side. Mentor organisations include Forethought, AI Policy Network and Coefficient Giving, plus independent mentors starting new orgs or scoping novel directions. A small number of fellows pursue independent strategy research (high bar) on topics such as macrostrategy for AI safety, prevention of concentration of power, exploration of better-futures scenarios, and implementing AI strategy in government or frontier AI companies. Includes weekly mentor and research-manager meetings, career support, startup incubation services, and access to the Constellation convening space (300+ weekly visits). Visa support: J-1; F-1/CPT (full-time only) or OPT, with DSO approval. Strongest applicants have deep familiarity with catastrophic AI risks, can communicate clearly, reason about complex issues and work autonomously. Preferred: macrostrategy/grantmaking/cause-prioritisation experience, ability to set and pursue research agendas, complementary career experience (policy, empirical AI), deep understanding of frontier AI companies / governments / investors. Application: resume + at least two references + highly recommended written samples (blog posts, papers). Stages: initial screening, mentor-specific work tests + reference check, interview. Deadline 3 May 2026, 23:59 Anywhere on Earth. Contact astra@constellation.org. Past Strategy fellows include co-founders of the AI Futures Project (authors of AI 2027) and people briefing senior international stakeholders on AI verification.
Fellowship for people with a solid technical or research background looking to shift their attention to AI safety and alignment. Not intended for beginners or casual explorers. Open to all nationalities. Scheduled to run 14 September 2026 to 5 February 2027.
Fully-funded nine-week research program for undergraduate, graduate and PhD students working on AI safety and alignment. Focus areas with AI Objectives Institute: AI political economy, gradual disempowerment, multi-agent institutional design. Programme covers interpretability, multi-agent safety, formal verification, civilisational resilience and more. Weekly 1-on-1 mentorship, 24/7 access to Harvard Yard offices. Particularly encouraged for students and researchers at Harvard, MIT and the wider Boston area, but open more broadly.
Top-tier track of the 2026 AWS Imagine Grant for US nonprofits with strong data practices, in the planning phase of incorporating frontier AI (generative AI, agentic AI, autonomous systems) as a core workload. Up to $200K cash + $100K AWS credits, plus hands-on project implementation support from the AWS Generative AI Innovation Center, training, and ongoing access. Two-round selection: Round One (open call) closes 5 June 2026; Round Two opens 10 August (invitation-only) and closes 14 September 2026; results mid-November; public announcement 1 December 2026. Eligibility: registered 501(c) nonprofits headquartered in the US. Educational institutions are explicitly excluded. NOTE: this is a nonprofit-only programme and is not open to individual artists, researchers, or for-profit studios.
Award track for US nonprofits running highly innovative cloud projects that leverage advanced services such as AI/ML, HPC, or IoT. Up to $150K cash + $100K AWS credits, plus AWS technical specialist guidance and training access. Round One closes 5 June 2026; Round Two by invitation, with full proposals 10 August to 14 September 2026. Eligibility: registered 501(c) nonprofits in the US. Educational institutions are excluded. NOTE: this is a nonprofit-only programme and is not open to individual artists, researchers, or for-profit studios.