AI2050 is a philanthropic initiative of Schmidt Sciences that funds exceptional researchers worldwide working on the hard problems required for AI to be hugely beneficial to society by 2050.
AI2050 is a philanthropic initiative of Schmidt Sciences that funds exceptional researchers worldwide working on the hard problems required for AI to be hugely beneficial to society by 2050.
People– no linked people
Updated 05/18/26Funding Details
Updated 05/18/26- Annual Budget
- $25,000,000
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- $125,000,000
Org Details
Updated 05/18/26AI2050 is a philanthropic research initiative housed within Schmidt Sciences, the science philanthropy founded in 2024 by Eric and Wendy Schmidt as a successor to Schmidt Futures. AI2050 was launched in February 2022 and is grounded in a central motivating question: 'It's 2050. AI has turned out to be hugely beneficial to society. What happened?' The program works backward from that aspirational outcome to identify the problems that must be solved and opportunities that must be seized for AI to benefit humanity. Eric and Wendy Schmidt committed $125 million over five years to support the initiative. Eric Schmidt and James Manyika serve as co-chairs, with Mark Greaves as Executive Director. The initiative operates through a closed nominations process, drawing on a network of approximately 300 AI experts worldwide to identify candidates. There is no open application process. Awards are made to both early-career and senior fellows for three-year research projects. In the fourth cohort announced in November 2025, 28 scholars received more than $18 million in fellowships (21 early-career fellows and 7 senior fellows). Across all cohorts, the program has supported 99 fellows at 42 institutions in 8 countries. Notable fellows have included Fei-Fei Li, Daniela Rus, David Autor, and Yejin Choi. In February 2025, Schmidt Sciences also announced a separate $10 million AI Science Safety program with 27 projects, including work by Yoshua Bengio on AI risk mitigation and Zico Kolter on AI attack vectors. The initiative has published a Working List of Ten Hard Problems in AI covering: (1) Capabilities, (2) Assurance, (3) Alignment, (4) Great Opportunities, (5) Economics, (6) Access, (7) Responsibility, (8) Geopolitics, (9) Governance, and (10) Meaning. This list is intended to evolve as AI capabilities progress. Fellows publish over 100 papers annually in leading journals and conferences, and 17 active cross-fellow collaborations are underway as of late 2025. Schmidt Sciences is headquartered in New York, NY, and AI2050 operates as one of its five programmatic centers alongside Astrophysics and Space, Biosciences, Climate, and Cross-Science programs.
Theory of Change
Updated 05/18/26AI2050 operates on the theory that identifying the right hard problems in AI and funding exceptional researchers to solve them is the most direct path to ensuring AI is beneficial by 2050. By investing in a global, interdisciplinary community of leading researchers at early and senior career stages, providing them with multi-year unrestricted support, and building connections among them, the initiative aims to accelerate progress on the technical, governance, and societal challenges that most threaten beneficial AI outcomes. Supporting open publication ensures that findings benefit the entire field rather than a single institution. The program explicitly addresses both capability advancement and safety/alignment challenges as complementary rather than competing concerns.
Grants Received– no grants recorded
Updated 05/18/26Projects– no linked projects
Updated 05/18/26Discussion
No comments yet. Be the first to share your thoughts.