Manifund

Theory of Change

Manifund's stated theory of change is to decentralize AI safety funding through regranting infrastructure. Rather than large institutional funders making all allocation decisions, Manifund delegates budgets of $100K+ to domain experts ("regrantors") who independently identify and fund early-stage projects. Austin Chen, the founder, frames this as "angel investing for charity":

"We want to see regrants that are like angel investments in undervalued people and projects... the regrants below feel more like post-Series A top-ups." -- Jesse Richardson, Manifund staff, April 2025

The mechanism is modeled on the FTX Future Fund's regranting program. Austin explicitly calls Manifund "Future Fund 2" and describes the appeal: "People were just empowered to have individual budgets of something ranging from hundreds of thousands to millions of dollars, where they could more or less make a decision on a grant without having to get external approval from committees." The key advantages over traditional funders (CG/OpenPhil, LTFF) are speed (<1 week to deploy), transparency (all grants and reasoning published), and hits-based giving (regrantors bet on undervalued opportunities rather than seeking committee consensus).

Manifund originally pursued a more ambitious theory: impact certificates ("venture funding for charitable endeavors"). This was their "most exciting project" in 2023, with backing from intellectual champions including Paul Christiano, Vitalik Buterin, and Scott Alexander. By mid-2024, they conceded impact certs had not found product-market fit: "we've learned a lot and are happy with the projects we've funded, but are less excited by impact certs than before -- it's been hard to get investor interest, and we still haven't found a use case where certs led to better funding decisions."

What They Do

Regranting (primary program): 2023: $1.8M distributed through 16 regrantors. 2024: ~$1.6M through 6 regrantors. 2025: $2.25M raised for 10 regrantors. Regrantors include researchers from DeepMind, Apollo Research, METR, Longview, Anthropic-adjacent figures. Grant sizes typically $5K-$50K, though some are much larger. Top cause area: Technical AI Safety (76% of 2023 funding by dollar amount).

Standout grants: First funding for Timaeus ($143K, accelerated DevInterp research by months), ChinaTalk ($37K, deep China-AI coverage ahead of DeepSeek surprise), Gavin Leech's AI safety review ($9K, induced follow-on from OpenPhil).

Grants that raised concerns internally: Epoch AI benchmark ($200K from single regrantor, too confidential, possible capabilities acceleration), CAIS/AI Frontiers ($250K from Dan Hendrycks to his own organization), SaferAI ($100K to org growing faster than its capacity).

Impact certificates (deprioritized): ACX Grants 2024 direct funding: $1.35M to 33 projects. The impact market attracted only ~$50K from investors. Micro-regranting experiment ($500 budgets to 24 people) was judged a "huge success" for engagement but tiny in scale.

Mox coworking space: SF's largest AI safety hub, launched Feb 2025. 40K sq ft at 1680 Mission. 183 members, 15 offices, 377 events in Year 1. Partners with MIRI, Redwood Research, GovAI, MATS, FAR.AI. Revenue ~$100K/month, expenses ~$130K/month. Seeking sustainability by end of 2026.

Fiscal sponsorship: Provides 501(c)(3) status to PauseAI and other organizations. Hosts ACX Grants on behalf of Scott Alexander.

Manifest conference (reputational liability): Annual event at Lighthaven (Lightcone Infrastructure), organized by Manifund staff. See "What Others Say" below.

Total grants trajectory: $2.06M (2023), $3.12M (2024), $2.25M raised (2025 regranting alone).

Key People

Austin Chen -- CEO, co-founder. Former Google SWE, Streamlit tech lead. Co-founded Manifold Markets (2021), left April 2024 to focus on Manifund full-time. Describes his North Star as "Winning." Self-described "perhaps the last SBF fanboy" who publicly argued SBF "should be freed and put back to work." Admits limited AI safety expertise: "My expertise is in startups and building websites and technology."

Rachel Weinberg -- Co-founder, engineer. Built initial platform from scratch in 2 weeks. Only Manifest organizer to express partial regret over the HBD speaker controversy.

Jesse Richardson -- Joined 2025. Former Mila - Quebec AI Institute. Authored the most analytically rigorous self-evaluation of Manifund's grants.

Team is approximately 3-4 people. Total compensation was $83,000 for the entire organization in 2023. Board of directors: Austin Chen, Barak Gila, Vishal Maini (Founding Partner at Mythos Ventures, a $14M AI fund; former DeepMind comms head).

Money and Incentives

Total budget: Operating costs are tiny. $83K total compensation in 2023. Most "revenue" ($3M in 2023) is flow-through donations passing from donors to grantees. Actual operating income comes from a 5% fee on donations (~$150K on $3M) plus SFF general support grants.

Revenue breakdown:

5% platform fee on donations: ~$100-150K/year
SFF general support grant: amount undisclosed, funded small regrantor budgets ($400K pool in 2023)
Individual donations: $95K from ~40 donors (2023)
Mox generates ~$100K/month revenue but costs ~$130K/month

Regranting funds (pass-through, not operating):

Anonymous donor "D": $1.5M/year (2023 and 2024) -- chose the large regrantors
2025: $2.25M raised from unspecified donors

Historical funding:

$500K seed from FTX Future Fund (2022) -- genesis of the 501(c)(3)
FTX clawback status: publicly unaddressed

Rejected funding: Applied to OpenPhil/Coefficient Giving, Lightspeed, and Y Combinator. All rejected. Manifund is the rare AI safety org that has never received CG/OpenPhil funding.

Key financial risks:

Extreme donor concentration: One anonymous person ("D") provides the bulk of regranting funds. If they walk away, the program collapses.
Operating sustainability is fragile: 5% fees on ~$3M in pass-through barely cover one salary. The org depends on SFF's continued goodwill for general support.
Mox is a large, separate financial commitment ($1.2M/year) that dwarfs the core grantmaking infrastructure cost.

Board COI: Vishal Maini runs a $14M AI venture fund while sitting on the board of an AI safety funder. No conflict-of-interest policy is documented.

Governance: Three-person board where the CEO holds a seat. No COI policy, no whistleblower policy, no formal grant evaluation process, no impact assessment. Austin acknowledged the board needs upgrading in January 2024; no changes have been publicly reported.

What Others Say

The Manifest/HBD controversy is the dominant external narrative. For three consecutive years (2023-2025), Manifund-organized Manifest conferences have platformed speakers with documented ties to scientific racism and the HBD movement, including Richard Hanania (published on HBDbooks.com, wrote for white supremacist outlets), Jonathan Anomaly (published "Defending eugenics"), Razib Khan (chased from NYT for racist remarks), Stephen Hsu (removed from MSU VP role), and Brian Chau (repeatedly made racist statements documented by journalists). After-parties hosted by neoreactionary figure Curtis Yarvin attracted many conference attendees.

Austin Chen's response: "I stand behind every one of the speakers we asked to come." Co-organizer Saul Munn: criticism has "done ~nothing to reduce the likelihood that i invite speakers who the commenters consider racist to the next manifest." Only co-organizer Rachel Weinberg expressed partial regret.

Community consequences: Peter Wildeford (Rethink Priorities co-CEO) quit Manifold entirely. Bob Jacobs (EA Ghent organizer) left the EA movement. Shakeel Hashim (former CEA communications head): "By far the most dismaying part of my work at CEA was the increasing realisation that a big chunk of the rationalist community is just straight up racist." Dustin Moskovitz signaled he would pull back from funding rationalist events, stating: "My view is that rationalists are the force that actively makes room for it."

SBF defense: Austin publicly argued SBF "should be freed and put back to work" and calls himself "perhaps the last SBF fanboy." His Manifold departure post listed SBF as an idol alongside Bill Gates. This is the CEO of a nonprofit that received its seed funding from FTX Future Fund and operates in a community devastated by SBF's fraud.

Positive assessment from Nuno Sempere (Manifund regrantor, disclosed COI): "Manifold is highly technologically competent... It seems very plausible to me that Manifund could do the same thing to CEA's Effective Altruism Funds: Create a product that is incomparably better."

David Thorstad (whose blog was itself Manifund-funded) wrote the most thorough critical account of the Manifest controversy, documenting each speaker's history with scientific racism in a 9,692-word analysis.

What's Absent

No 2024 annual review despite it being Manifund's biggest year
No conflict-of-interest policy despite documented self-dealing (Dan Hendrycks granting to own org)
No grant outcome tracking or impact evaluation system
No board expansion since 2024 acknowledgment of need
Identity of largest funder ("D," $1.5M/year) remains unknown
FTX clawback status never publicly addressed
No formal separation between Manifund and the Manifest conference
No independent evaluation comparing Manifund's regranting to LTFF, SFF, or CG
OpenPhil/Coefficient Giving has never funded Manifund