Projects

Tools, collectives, and initiatives advancing AI redistribution. These are the efforts worth knowing about — and supporting.

How does open AI get funded?

Open AI projects need sustainable funding models. Understanding these helps you evaluate projects and find ways to contribute:

•
Foundation-funded — Nonprofit research, grant-supported
•
Volunteer collective — Researchers contributing time and resources
•
Open core — Open source + paid services
•
AI-bootstrapped — Small teams using AI to build at scale
•
Public compute — Government-funded infrastructure
•
Mission-driven startup — VC-funded but values-led
•
Protocol + ecosystem — Open standard, value in adoption
Filter by:

15 projects

Foundation-funded

⚠ Grant-dependent; funding cycles create uncertainty

policyresearchadvocacy

Foundation-funded

⚠ Small organization; capacity limits scope

advocacyresearchpolicy
BLOOMactive

Public compute

⚠ Model now dated relative to frontier; no funded update path

modelresearchdata

Foundation-funded

⚠ Close ties to AI labs may limit critical independence

policyresearchadvocacy

Foundation-funded

⚠ Copyright status of crawled content legally contested

datainfrastructureresearch

Foundation-funded

⚠ Grant-dependent; long-term funding uncertain

researchadvocacypolicy

Volunteer collective

⚠ Volunteer-dependent; burnout and turnover risk

researchmodeldata

Open core

⚠ VC-funded ($395M+); growth pressure may conflict with openness

techmodeldata
Lelapa AIactive

Mission-driven startup

⚠ VC-funded; investor returns may conflict with mission

techmodelresearch
Masakhaneactive

Volunteer collective

⚠ Volunteer-dependent; sustainability relies on researcher time

researchdatamodel

Protocol + ecosystem

⚠ Anthropic-originated; governance not yet community-controlled

standardsinfrastructuretech

Foundation-funded

⚠ Mozilla's financial challenges affect project stability

dataresearchinfrastructure
Ollamaactive

Open core

⚠ VC-funded; long-term business model unclear

techinfrastructure

Open core

⚠ Code execution is powerful but risky; security concerns

techinfrastructure
WebLLMactive

AI-bootstrapped

⚠ Early stage; adoption and ecosystem still building

techinfrastructurestandards