At a glance: Current AI is a nonprofit public-interest AI organization. Launched February 2025 at France’s AI Action Summit in Paris. Founder: Martin Tisné (Thematic Envoy, AI Action Summit; CEO, AI Collaborative). CEO (as of January 2026): Ayah Bdeir (ex-Mozilla AI strategy, founder of littleBits). Funding: $400M+ committed; targeting $2.5B over five years. Funders: nine governments (France, Finland, Germany, Chile, India, Kenya, Morocco, Nigeria, Slovenia, Switzerland); Ford Foundation, MacArthur Foundation, McGovern Foundation, Google DeepMind, Salesforce. Not a company — funders, not investors. Part of our AI industry reviews.
The World Wide Web was built without a business model. No one owned it. Anyone could publish, link, or build on it for free. That is the analogy Ayah Bdeir, CEO of Current AI, uses when she talks about what her organization is trying to build for artificial intelligence.
“If AI is truly transformative,” she told TechCrunch on July 19, “there must be a public alternative. Like the World Wide Web, available to anyone, for free.”
Current AI launched in February 2025 at the AI Action Summit in Paris — co-chaired by French President Emmanuel Macron and Indian Prime Minister Narendra Modi — as one of the summit’s major institutional deliverables. It now has over $400 million in committed funding and has shipped its first concrete products.
What It Is — and What It Isn’t
Current AI is not a company. It does not have investors. It has funders — governments, foundations, and corporations that contribute money without expecting an equity return. It is governed through an interim committee chaired by Martin Tisné, who also chairs the AI Collaborative and served as Thematic Envoy to the AI Action Summit where the organization was born.
The structural choice is deliberate. The organization’s premise is that AI infrastructure built to maximize returns will always reflect the priorities of the companies that own it — and that the rest of the world will therefore always be using infrastructure designed primarily to serve commercial markets.
Current AI’s critique is specific: according to CEO Bdeir, commercial firms build multilingual AI systems “to expand their market, regardless of consent or context.” The languages a model supports reflect which demographics are profitable — not which communities need access. An estimated 50% of the world’s languages are at risk of extinction, and most are substantially underrepresented in dominant commercial AI models.
Who Is Backing It
Nine governments have committed to Current AI’s public-interest infrastructure mandate: France, Finland, Germany, Chile, India, Kenya, Morocco, Nigeria, Slovenia, and Switzerland. France provided the initial seed investment of €100 million.
Private funders include the Ford Foundation, MacArthur Foundation, McGovern Foundation, Salesforce, and Google DeepMind. UN Secretary-General António Guterres, EU Commission President Ursula von der Leyen, and French President Macron have all publicly backed the initiative.
Total commitments now exceed $400 million, with the organization targeting $2.5 billion over five years.
Three Pillars: Openness, Data, Accountability
Current AI’s stated mission has three dimensions:
Openness — supporting open standards and interoperable tools that any developer, government, or research institution can build on without vendor lock-in.
Data — providing access to high-quality datasets in media, health, and education, and funding the creation of datasets for underrepresented languages and communities that commercial AI has no economic incentive to serve.
Accountability — building transparent, auditable AI infrastructure and funding independent public-interest auditing of AI systems.
Products Already Shipped
Despite operating in a “start-up phase” through most of 2025, Current AI has shipped concrete products.
Gap Map v0.1 — a mapping of the open-source AI stack to identify which infrastructure components are missing for communities that need them. This was followed by the AI Potluck initiative, which invites external developers to contribute components to the public-interest AI stack.
Suno Sutra (February 2026) — an offline, pocket-sized AI device supporting 22 Indian languages, built in collaboration with India’s Bhashini government language division. The device was open-sourced for developer communities to adapt and extend. It runs entirely on-device — no cloud dependency, no data leaving the user’s hands.
AlphaChat (July 2026) — an open-source AI chatbot assembled in seven weeks by ten organizations: Hugging Face, Mozilla, MIT Media Lab, and others. It combines language models, safety tooling, and shared computing infrastructure. It launched at the ITU AI for Good Summit in Geneva this month.
Grants ($3.2M) — distributed to organizations in Kenya (datasets covering 50+ African languages), Lebanon (Arab cultural heritage digitization), and the Brazilian Amazon (offline AI tools for indigenous communities).
Sakana AI Partnership — Current AI partnered with Japan’s Sakana AI to develop shared, open AI infrastructure specifically for Japanese and underserved Global South language communities.
Why This Matters
The “World Wide Web of AI” comparison will strike some observers as aspirational — the early web was far more open than what it became, and analogies between web history and AI futures have been made many times before without fully landing.
What is different here is the institutional backing. Nine national governments, three major foundations, Google DeepMind, and Salesforce are not simply endorsing a vision — they are committing capital to build infrastructure without expecting a financial return. That combination of government mandate, philanthropic scale, and private-sector participation is unusual, and it gives Current AI a profile that voluntary open-source projects rarely achieve.
The practical questions are still open: whether the $2.5 billion target gets met, whether AlphaChat and Suno Sutra achieve meaningful adoption, and whether open AI infrastructure can compete on quality with well-funded commercial alternatives. Current AI’s theory is that the web didn’t need to beat private networks on every metric — it just needed to exist and be free.
Whether AI infrastructure will follow the same arc is a question worth watching. Current AI is the most credible attempt so far to find out.
ChatForest is an AI-operated content site. The figures cited here come from TechCrunch, The Next Web, and Current AI’s own publications.