TECHNOLOGY · VERIFIED DEVELOPMENT
Cloudflare Launches Two European Open AI Models on Workers AI
WHY IT MATTERS
The launch of these open AI models promotes national resilience and sovereignty, providing decentralized and affordable access to AI. It also highlights the importance of choice and freedom in AI access, which is critical for governments and organizations to build effective AI defenses.
What happened
Cloudflare has announced the launch of two European open AI models, EuroLLM and Apertus, on its Workers AI platform. EuroLLM supports 35 languages, including all 24 official EU languages, and was developed with support from Horizon Europe and the European Research Council.
Apertus is Switzerland's first large-scale, fully open, multilingual language model, trained on over 15 trillion tokens across 1,500 languages. Both models are designed to provide decentralized and affordable access to AI, promoting national resilience and sovereignty.
Cloudflare's security team has also built AI defenses that work with any model, preventing vendor lock-in and ensuring that losing access to one model doesn't compromise defenses. The launch of these models follows Cloudflare's argument for choice and freedom in AI access, which has been hardened by the use of frontier models in cyber attacks and the restriction of access to some models.
PRIMARY SOURCES
One year later: Sovereign AI and the fight for choice
Cloudflare (company statement) · Carly Ramsey · Corporate primary source; facts only, no copied text or images; link and attribution required
CORRECTIONS & UPDATES
- Revision 1 · Initial ingestion · Oct 1, 2026, 1:45 PM
- Revision 2 · Source update detected · Oct 1, 2026, 1:45 PM