Current AI, a Paris-seeded nonprofit founded in February 2025, has assembled $400M in committed funding to build public AI infrastructure as an alternative to systems from OpenAI, Google, and Anthropic. The French government put in the first $100M, with Ford Foundation, MacArthur Foundation, DeepMind, and Salesforce filling out the rest. Last month, Current AI deployed its first $3.2M in grants across four organizations in Kenya, Lebanon, and the Brazilian Amazon.
The pitch is that AI's dominant systems belong to private companies headquartered in a handful of Western cities, and that a farmer in rural India photographing a dying plant should not need English — or a Silicon Valley account — to get an answer. In February 2026 at the India AI Summit, Current AI teamed up with Bhashini, the Indian government's AI language division, to ship Suno Sutra: a pocket-sized offline device running AI in 22 Indian languages, fully open-sourced for developer communities to build on.
“In India, there are hundreds of different languages and dialects, and right now AI is not representing them”— Ayah Bdeir, Current AI CEO
CEO Ayah Bdeir joined in January after running Mozilla's AI strategy. She previously founded littleBits, the STEM education company that sold to Sphero in 2019. Current AI's founder, Martin Tisne, structured the organization as a public-private partnership pulling from governments, corporates, and philanthropies rather than venture capital.
Key facts
- 01Current AI has secured $400M in total committed funding, seeded by $100M from the French government and joined by Ford Foundation, MacArthur Foundation, DeepMind, and Salesforce.
- 02The nonprofit deployed $3.2M in grants last month to four organizations across Kenya, Lebanon, and the Brazilian Amazon.
- 03Its Suno Sutra device, built with India's Bhashini, runs AI in 22 Indian languages offline, no internet required.
- 04Alpha Chat, an open-source chatbot, was assembled in seven weeks by a coalition of ten organizations including Hugging Face, Mozilla, and MIT Media Lab.
- 05CEO Ayah Bdeir, formerly Mozilla's AI strategy lead and littleBits founder, joined in January 2026.
The distinction matters for how the money moves. Bdeir told TechCrunch that the backers "are funders, not investors," which means no equity stake, no return timeline, and no pressure to monetize the models or the datasets. That structural choice is the point: the models, data, and tools are meant to remain accessible without gating.
The grant portfolio shows the scope. In Kenya, Masakhane is building AI datasets across 50+ African languages for health, farming, and education. Lebanon's Institute for Worldmaking is digitizing Arab cultural history into machine-readable databases that stay under community control. Portal sem Porteiras is building offline AI tools with Indigenous Amazon communities, keeping the data inside the territory. Kenya's African Internet Rights Alliance is developing audit tools for AI accountability across the continent.
Earlier this month in Geneva, Current AI launched Alpha Chat, an open-source chatbot assembled in seven weeks by a coalition of ten organizations including Hugging Face, Mozilla, and MIT Media Lab. Each contributor brought part of the stack — a language model, safety tooling, compute. Current AI also announced a partnership with Tokyo-based Sakana AI to build a shared open-source stack supporting Japanese and communities across the Global South.
Bdeir frames the ownership question bluntly. She argues that multilingual expansion by large model providers happens "regardless of consent or context," and that missionary Bible translations frequently become training data for Indigenous languages before the communities involved have set any rules. Current AI's approach is to store models and data locally, bring in community experts before anything is built, and write consent protocols into the pipeline so communities can halt the process.
None of the grantees has fully solved data ownership, and $3.2M split four ways is a fraction of what a single frontier training run costs at OpenAI or Google. Half the world's spoken languages face extinction, and the gap between what a well-funded lab can build and what a distributed coalition of nonprofits can build widens with every compute-scaling cycle. Bdeir's counter is that scale is the wrong metric — that a tool built in Kenya usable by an elder in the Amazon is a different kind of win than a benchmark score.
For the AI market, Current AI is the first credibly capitalized attempt to treat foundation-model infrastructure the way the early web was treated: as public utility rather than proprietary asset. Whether $400M can hold that line against private labs spending that much on a single training run is the open question. But the political geometry — French state capital, US philanthropic capital, DeepMind and Salesforce lending resources to a nonprofit competitor to their own commercial products — is unusual enough that the experiment will be watched closely by every government now writing an AI sovereignty strategy.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
