Current AI Is Building a $400M Open Alternative to Big Tech's AI Models
Nonprofit Current AI, backed by $400 million in committed funding, is developing open-source AI infrastructure to protect linguistic diversity and local data sovereignty from consolidated corporate control.
Current AI, a nonprofit founded in February 2025 by Martin Tisne, is building a public alternative to the proprietary AI systems dominated by a handful of Silicon Valley corporations. Structured as a public-private partnership, the organization draws support from governments, philanthropic bodies, and corporate donors. Its $400 million in committed funding includes a $100 million seed contribution from the French government, with additional backing from the Ford Foundation, the MacArthur Foundation, Google DeepMind, and Salesforce.
CEO Ayah Bdeir, who previously led AI strategy at Mozilla and founded education technology startup littleBits, is driving the organization's operations. Her argument against the status quo is direct: the dominant AI models are controlled by a small set of private companies, primarily OpenAI, Google, and Anthropic, that build multilingual systems to expand their market reach rather than to serve local communities. Cultural context and community consent are rarely part of the equation. Without a credible public alternative, large portions of the world's linguistic and cultural diversity risk being sidelined entirely.
Current AI is already shipping products. In India, the organization worked with Bhashini, the government's language division, to produce Suno Sutra, an open-source, pocket-sized device that operates entirely offline and provides translation and dialogue across 22 Indian languages. At the AI for Good Summit in Geneva, the nonprofit also released Alpha Chat, an open-source chatbot assembled in seven weeks by a coalition of ten organizations including Hugging Face, Mozilla, and the MIT Media Lab.
Data sovereignty sits at the center of the nonprofit's strategy. Bdeir argues that communities must maintain ownership and control over their own records, memories, and traditions, rather than having foreign tech companies extract that data for model training. To back that position operationally, Current AI has built consent protocols that allow local experts to pause or stop data pipelines during development, giving communities a meaningful check on how their information is used.
The organization is also partnering with Tokyo-based startup Sakana AI to develop a shared open-source stack. The collaboration targets Japanese language models while simultaneously producing tools for underrepresented communities across the Global South, with the aim of building a decentralized network where improvements are shared across all participants.
What This Means for Africa
Africa is a direct focus, not an afterthought. A significant portion of Current AI's initial $3.2 million grant cohort is flowing to the continent. Masakhane, a grassroots research community based in Kenya, is receiving funding to build datasets covering more than 50 African languages across healthcare, agriculture, and education. The African Internet Rights Alliance is using Current AI grants to develop auditing tools designed to hold global AI systems accountable within African markets. Both efforts target a gap that African developers and startups know well: Western models are expensive, culturally imprecise, and built without African input. Sovereign datasets give African institutions a foundation to build tools calibrated to local economic realities, without depending on platforms that were never designed with them in mind.
Real technological access means giving every culture the tools to build its own digital future, not just a translated version of someone else's.
Source: TechCrunch
Written by
StacyAI-assisted news curation. Every story is reviewed by our editors before publication.
