Q.ANT Opens Photonic AI Hardware Pilot on IONOS Bare Metal
What happened
On November 9, 2026, Q.ANT began a closed beta pilot program that lets selected users run their own AI workloads on its second-generation photonic processors, hosted on IONOS bare-metal servers in Germany. The pilot is free and targets participants with specific use cases and the technical ability to manage a bare-metal server. This initiative is designed to test the commercial viability of Q.ANT's photonic hardware for a range of AI tasks. The program follows an earlier commercial agreement between Q.ANT and IONOS, with the goal of moving from software simulations to real hardware evaluations.
Key details
- Pilot launched November 9, 2026, for select users in Germany.
- Runs on second-generation photonic Native Processing Units (NPUs) via IONOS bare-metal servers.
- Supports AI tasks including industrial vision, time-series forecasting, and large-model inference.
- Pilot participants must have a concrete use case and server administration skills.
- Independent tests reported up to 50x speedup and 25x acceleration in image recognition, with 50–84% lower power use compared to the previous generation.
Why it matters
Companies and research groups with advanced AI needs can now test photonic hardware for their own workloads, potentially reducing the number of parameters and energy required for certain tasks. This gives these users a direct way to evaluate whether Q.ANT's technology can improve efficiency or performance in their production environments. The pilot's results will help determine if and how photonic co-processors can be integrated alongside CPUs and GPUs in commercial AI infrastructure.