Hetzner Inference: First Look
Summary
Sliplane's analysis covers Hetzner's experimentation with an OpenAI-compatible LLM inference API running on Hetzner infrastructure. The article emphasizes that the offering is experimental (no billing, no SLA, one model) and discusses potential business and hardware implications, including the viability of OpenWeight-style inference and the economics of running GPUs in Hetzner data centers.