Hetzner launches AI inference API in Europe
Qwen 3.6-35B (262k context), 153ms/first token, 224 tokens/sec, Fully Compatible with OpenAI API. Hetzner, one of Europe’s leading web hosting companies, has released an experimental AI Inference API.
Hetzner, one of Europe’s leading web hosting companies, has released an experimental AI Inference API. They’re running Qwen 3.6-35B (262,000 tokens of context, 153 ms to first token, 224 tokens per second) which is fully compatible with the OpenAI API. There are no bills or SLAs associated with this service at the moment, but it’s an early experiment.
European Web Hosting Companies are Starting to Build Their Own AI Infrastructure. This will be a direct competitor to expensive US Cloud Services such as Azure and AWS. If Hetzner starts deploying server-grade GPU’s like the Nvidia B200 instead of using consumer GPUs, the price of AI Inference in Europe will fall significantly and the US cloud providers will lose their monopoly on GPU resources.
Source: sliplane.io
Free course
Stop reading about AI — start building with it
The free Claude Code course: your first site, tool or game — no coding. No upsells, no cross-sells — nothing to buy here.
Start free →
Author
Evgenii Arsentev
PhD · Chief Executive Officer, digital health
Articles · Latest articles