Funded through philanthropic grants, sovereign research partnerships, and selective licensing — not venture capital.
alphabell does not raise conventional venture capital. The lab is funded through a mixture of philanthropic grants, sovereign research partnerships, and revenue from selective licensing of derivative tooling. No equity is issued. No funder may unilaterally redirect a research thread. Every funder is approved as a signed proposal.
The bet behind not taking VC
Conventional venture capital is structurally incompatible with the lab's commitments. VC requires equity, which we do not issue. VC requires growth, which we are not particularly trying to optimize for. VC requires exits, of which the natural ones for an AI lab are acquisition or IPO, both of which would consolidate the lab into a structure that contradicts its decentralization commitments. And — perhaps most importantly — VC sets up funder incentives that are well known to interfere with publication independence and research-direction independence over the medium term.
Not taking VC has costs. It means we have to assemble funding from many smaller sources, none of which alone can support the lab. It means we have to spend time on funder relationships that a VC-backed lab would not. It means our fundraising cycles are 12 to 18 months and structurally lumpy. We do not pretend the costs are zero; we believe the alternative is worse.
The three funding sources
Philanthropic grants
The largest single category. About half of the lab's operating budget in 2025 came from philanthropic grants — primarily from a consortium of foundations interested in long-horizon AI research, AI safety, and the institutional capacity for science outside of corporate research. Grants are typically structured as multi-year commitments with annual tranches.
The grant terms are documented and standardized: no scope-of-research conditions, full publication independence, audit rights for the funder limited to compute usage and capability-evaluation summaries. We have, on three occasions in the lab's history, declined a philanthropic grant whose terms would have required scope conditions we were not willing to accept.
Sovereign research partnerships
About 38% of the 2025 budget came from sovereign research partnerships — typically academic clusters or national research-funding bodies. Sovereign partnerships are approved at a higher governance threshold than philanthropic grants and include a counterparty-due-diligence step that screens for jurisdictional research-export controls and any commitments inconsistent with the charter.
Active sovereign partners are listed in /compute. The full counterparty list is reviewed annually by the long-tenured-contributor quorum.
Licensing of derivative tooling
About 12% of the 2025 budget came from selective licensing of tooling derived from alphabell research. The tooling licences are non-exclusive, time-limited, and structured to ensure that licensees do not acquire any influence over the lab's research direction. Open-source releases of the underlying tooling continue regardless of licensing.
Licensing partners are listed in /press as part of the lab's public counterparty record.
Audit and transparency
The lab publishes a high-level financial summary every year, in the public news index. The 2024 summary is at news/funding-summary-2024 (forthcoming); the 2025 summary will be published in 2026-Q1.
The summary contains: total operating budget, breakdown by source category, breakdown of expense (compute / contributor stipends / anchor operating costs / governance overhead), and any unusual one-off items. It does not contain per-funder amounts; funders are listed by category and approved status.
What contributors are paid
Contributors are not employees. Contributors receive stipends from the lab during periods of active cell participation. The stipend is set by the long-tenured-contributor quorum and is independent of country of residence; the lab does not negotiate stipends individually. The 2025 stipend is approximately $9,400/month for active full-time contributors, prorated for partial-time participation.
The stipend is funded out of the lab's operating budget. It is not equity. It is not bonus-eligible. Contributors who want to be paid more than the stipend, or to receive equity, work elsewhere — and we wish them well. This is one of the lab's load-bearing trade-offs.
What we will not take
- Funding from a single source large enough to redirect the lab. No funder represents more than 14% of operating budget.
- Funding with scope-of-research conditions. Three declined in the lab's history.
- Equity. The lab does not issue equity.
- Funding tied to public endorsement of a funder's product, position, or strategy. See charter Article 5.
- Anonymous funding above a small threshold. Funders above $250k/year are publicly identified.