Current AI Raises $400M to Build Public AI Infrastructure
by Vivek Gupta - 1 day ago - 4 min read
Current AI, a nonprofit backed by governments, technology companies and philanthropic organisations, is building an open AI ecosystem designed to function more like the early World Wide Web than a closed commercial platform.
The organisation wants to connect public datasets, open models, safety tools, computing infrastructure and community-led applications through shared standards. Its aim is to reduce dependence on a small group of companies that currently control many of the most advanced AI systems.
Founded in February 2025 by technology policy expert Martin Tisné, Current AI has secured more than $400 million in committed support. France contributed an initial €100 million, while supporters include the Ford Foundation, MacArthur Foundation, Google DeepMind and Salesforce. The nonprofit plans to mobilise $2.5 billion over five years.
A Public Alternative to Closed AI Platforms
Current AI is not trying to build a single chatbot to compete directly with OpenAI, Google or Anthropic. Instead, it is developing and connecting different layers of the AI ecosystem.
The organisation’s projects focus on open datasets, multilingual models, transparent development, local data governance and tools that can be adapted by governments, researchers and communities.
| Priority | Current AI’s approach |
|---|---|
| Open infrastructure | Shared datasets, models and development tools |
| Language inclusion | Support for underrepresented languages |
| Community control | Local ownership of sensitive data |
| Transparency | Auditable models and development processes |
| Global access | Tools designed for public and nonprofit use |
The project addresses a growing concentration problem in AI development. Stanford University’s 2026 AI Index found that industry produced more than 90% of notable AI models released in 2025, while leading systems often disclosed limited information about training data, model size and development methods.
Offline AI for 22 Indian Languages
One of Current AI’s main projects is Suno Sutra, a pocket-sized offline AI device developed with Bhashini, the Indian government’s digital language initiative.
The device can run AI services in 22 Indian languages without a continuous internet connection. It could allow farmers to analyse crop problems, health workers to access translated information and communities to document local knowledge without uploading sensitive data to external servers.
Suno Sutra was presented as a prototype in February 2026 and released as open-source technology that developers can adapt.
The project reflects Current AI’s wider goal of supporting languages that are poorly represented in the datasets used to train commercial AI systems.
Community-Led AI Projects Receive $3.2 Million
In June 2026, Current AI distributed $3.2 million to four organisations working in Africa, Lebanon and the Brazilian Amazon.
| Organisation | Focus |
|---|---|
| Masakhane | AI datasets for more than 50 African languages |
| Institute for Worldmaking | Digital archives of Arab cultural history |
| Portal sem Porteiras | Offline AI for Indigenous Amazon communities |
| African Internet Rights Alliance | AI auditing and accountability tools |
These projects are designed around community participation and local control. Current AI argues that publicly available cultural or language data should not automatically be treated as freely usable without consent.
Alpha Chat Tests a Shared Open AI Stack
Current AI also launched Alpha Chat in July 2026 at the AI for Good Summit in Geneva.
The free, open-source chatbot was assembled in about seven weeks by ten organisations, including Hugging Face, Mozilla and the MIT Media Lab. Different partners contributed the base model, interface, safety tools and computing resources.
Alpha Chat is not positioned as a direct performance rival to the largest commercial chatbots. Its purpose is to show that multiple organisations can combine open components to create a usable AI service.
This distributed approach is central to Current AI’s strategy. Rather than allowing one company to control the full stack, separate organisations can build compatible datasets, models, applications and safety systems.
Funding Remains the Biggest Test
Current AI’s ambitions are significant, but its funding remains small compared with the billions being invested in commercial AI infrastructure.
Training advanced models requires expensive chips, data centres, energy and specialist talent. The nonprofit must also establish clear standards around open-source licensing, data consent and community ownership.
Its success will therefore depend less on producing the most powerful model and more on whether it can make open AI infrastructure practical, trusted and widely usable.
Current AI’s early projects suggest an alternative direction for the industry: AI built as shared infrastructure rather than a collection of closed platforms. The challenge now is proving that this model can scale.