$1 billion is not enough; OpenAI Foundation must start spending tens of billions each year
webAuthor
Credibility Rating
Good quality. Reputable source with community review or editorial standards, but less rigorous than peer-reviewed venues.
Rating inherited from publication venue: LessWrong
A LessWrong opinion piece critiquing the scale of OpenAI's safety funding relative to its capabilities work, relevant to debates about AI lab governance and whether safety investment is keeping pace with AI development.
Forum Post Details
Metadata
Summary
This LessWrong post argues that OpenAI's nonprofit foundation is drastically underfunding AI safety relative to the scale of AI capabilities development, and calls for the foundation to dramatically increase annual safety spending to tens of billions of dollars. The author contends that current safety investment is insufficient given the pace of AI progress and the magnitude of potential risks.
Key Points
- •Current AI safety spending by OpenAI is orders of magnitude too small relative to capabilities investment and the stakes involved.
- •The OpenAI Foundation should be spending tens of billions annually on safety research, not the current ~$1 billion level.
- •The disparity between capabilities and safety funding represents a structural governance failure within the organization.
- •Increased safety spending is framed as a moral and existential imperative given the potential for transformative AI systems.
- •The argument implies broader critique of AI lab governance and the need for accountability mechanisms to ensure safety-capabilities balance.
Cached Content Preview
# $1 billion is not enough; OpenAI Foundation must start spending tens of billions each year By Davidmanheim Published: 2026-03-25 OpenAI is now a public benefit corporation, with a charter that demands they use AGI for the benefit of all, and do so safely. To justify this structure to the Attorneys General of Delaware and California, they split off the nonprofit OpenAI Foundation, and instead of full ownership, gave it 27% equity, worth well over $150 billion - what some have called [the largest theft in human history](https://thezvi.substack.com/p/openai-moves-to-complete-potentially). They said at the time they would donate a small fraction of the value, $25 billion, eventually. This week, [they announced a team, and a commitment](https://openaifoundation.org/news/update-on-the-openai-foundation); giving "at least $1 billion" over the coming year; [in the full post](https://forum.effectivealtruism.org/posts/ME7yYXondbFpTCrri/usd1-billion-is-not-enough-openai-foundation-must-start), I argue that given their values and expectations, everyone should agree that this is far too little, moving far too slowly.
14ae268bacabc883 | Stable ID: sid_GsP1jV9CJR