Press Release
Launch of project advances AI infrastructure available to the public through strategic collaboration with leading European research institutions
2 September, 2025 – A new global initiative, the Public AI Inference Utility, officially launches today to provide international access to Apertus, the groundbreaking open-source language model developed by the Swiss AI Initiative. This strategic collaboration represents a significant step toward building trustworthy, public AI infrastructure that serves the global public interest.
Apertus represents a new generation of open models built from the ground up around democratic values: transparency, safety, and accessibility. Developed through a rigorous, publicly accountable process, Apertus has earned international recognition for its technical excellence and commitment to open, trustworthy governance.
The Public AI Inference Utility is an infrastructure layer that enables global users—especially in under-resourced and non-commercial settings—to access and use Apertus. The Utility coordinates donated compute resources and institutional support from partners around the world, expanding access to high-quality AI tools beyond traditional commercial channels.
The collaboration will enable broader international deployment of Apertus while establishing new frameworks for public AI governance, procurement, and institutional stewardship. The project is supported by donated compute from global partners—including Amazon Web Services, AI Singapore, Exoscale, the Swiss National Supercomputing Centre (CSCS), Cudo Compute, National Computational Infrastructure Australia, and many others—that will allow people outside of Switzerland to access the model. Essentially, it provides technical infrastructure and institutional support to help the public worldwide access Apertus.
“The Utility is a clear example of what it means to treat AI as public infrastructure,” said Joshua Tan, lead maintainer of the Public AI Inference Utility. “Instead of waiting for the private sector to set the terms, we are helping public institutions take the lead. Apertus shows how countries can work together to build more accessible AI systems.”
The launch of the Public AI Inference Utility is part of a broader wave of initiatives across Europe, Asia, and the Americas to reimagine the future of public and sovereign AI. The effort is supported by Metagov, a nonprofit research lab that has helped lead the movement for public AI—a vision of AI as public infrastructure like highways, water, or electricity.
The Apertus deployment will help establish new frameworks for public procurement, institutional stewardship, and international cooperation around AI infrastructure. It also offers a living demonstration of how open models, when paired with public institutions, can offer a viable alternative to the corporate monopolies that currently dominate the AI landscape.
About Metagov
Metagov is a nonprofit research lab. We engineer institutions. Originally created as a collaboration between Harvard’s Berkman Klein Center and the MIT Media Lab, Metagov officially spun out as an independent nonprofit in January 2020. The Public AI Inference Utility team was supported through grants from Mozilla, the Future of Life Institute, and the Center for Cultural Innovation. For more information, visit metagov.org.
Media Contact:
Katherine Gorman, Communications Lead
katherine@metagov.org
+1 (410) 599-2012
About Apertus and the Swiss AI Initiative
Apertus was developed as part of the Swiss AI Initiative, led by EPFL and ETH Zurich. It is the result of a collaborative effort bringing together researchers, engineers, and students from across Switzerland, alongside the engineers and infrastructure of the Swiss National Supercomputing Centre (CSCS). This collective expertise, spanning multiple institutions and disciplines, has made the development of Apertus possible.