Current AI, a nonprofit founded in February 2025, is trying to build public AI infrastructure that functions the way the early World Wide Web did: open, free, and not owned by any single company. Its most visible product so far, Suno Sutra, is a pocket-sized offline device built with India’s government language division that runs AI in 22 Indian languages without needing an internet connection at all. A nonprofit shipping working hardware for a problem no major AI lab has prioritized is what NewsTrackerToday opens onto as the more concrete story than the mission statement alone.
CEO Ayah Bdeir, who joined in January after leading Mozilla’s AI strategy and previously founded STEM company littleBits, frames the core problem in blunt terms: every major AI system today, from the largest U.S. labs to their closest rivals, belongs to a private company. Half the world’s spoken languages face extinction, and English-dominated training data means a large share of the world’s languages, and the cultures carried inside them, get left out of the systems increasingly mediating daily life.
Isabella Moretti reads the funding structure that makes this nonprofit different from a typical AI startup: “The French government seeded Current AI with $100 million, joined by the Ford Foundation, MacArthur Foundation, and two major tech companies, bringing total committed funding to $400 million. Bdeir is explicit that these are funders, not investors, which changes the entire incentive structure. There’s no expectation of financial return driving product decisions here, which is precisely what lets Current AI prioritize a device for 22 Indian languages over a mainstream feature aimed at the largest possible market.” That funder-not-investor structure, more than the dollar total, is what NewsTrackerToday weighs against the incentives shaping every other AI lab’s roadmap.
The organization’s grant-making moves in small, deliberate increments rather than chasing scale. Last month it allocated $3.2 million across four organizations: a Kenyan project building AI datasets across more than 50 African languages, a Lebanese institute digitizing Arab cultural history into community-controlled databases, a Brazilian effort building offline AI tools with Indigenous Amazon communities, and a Kenyan audit initiative holding AI systems accountable across the continent.
Sophie Leclerc, who covers the technology sector, reads the technical architecture choice as the more interesting long-term bet: “Current AI just launched Alpha Chat, an open-source chatbot built in seven weeks by a coalition that includes Hugging Face, Mozilla, and MIT Media Lab, each contributing a different piece of the stack, a language model, safety tooling, computing power. That distributed-contribution model is a genuinely different way to build AI infrastructure than a single company assembling everything in-house. Whether it can match the pace and polish of a well-funded private lab is the real open question, but the coalition structure itself is worth watching regardless of the answer.” That coalition model, more than the chatbot’s specific capabilities, is what News Tracker Today folds into as the actual experiment underway here.
A newer partnership extends this further: Current AI struck a deal with Sakana AI, a Tokyo-based startup focused on what it calls sovereign AI, to build a shared open-source stack supporting Japanese language and culture alongside underserved communities across the Global South. Bdeir has been direct about how this differs from Big Tech’s own multilingual efforts, which she describes as expanding market reach regardless of consent or context, sometimes training on missionary Bible translations before the communities whose languages they represent have set any rules for how their own data gets used.
None of Current AI’s individual grantees have fully solved the problem of building AI that genuinely serves a small, specific community rather than defaulting to scale. Bdeir treats that as the point rather than a shortcoming, arguing that building the ownership question directly into each project’s design matters more than the size of any single budget. Whether a $400 million public-private nonprofit can meaningfully compete with AI labs spending that same amount in a single week, or whether Current AI’s real value is proving a different model exists at all regardless of its eventual scale, is what NewsTrackerToday settles on as the actual test facing this organization over the next several years.
