Q. ANT launches open-source SDK for photonic computing

Stuttgart-based startup Q. ANT has released the world’s first open-source software development kit (SDK) for photonic computing. The Q. ANT Native Computing Toolkit, available on GitHub, allows developers to write, run, and test applications using Python or C that are designed to run on photonic processors rather than traditional electronic ones.

The SDK includes an API, documentation, and a simulation backend that enables programmers to validate applications on standard CPUs before deploying them to actual photonic hardware. This approach aims to build a developer ecosystem similar to Nvidia’s CUDA model. By performing computations in the optical domain, Q. ANT’s chips aim to reduce the energy consumption typically required to shuttle data between memory and processors in classical AI computing.

While the current toolkit focuses on running pre-trained models—such as those for object identification and handwriting recognition—full hardware access is expected in the coming months via cloud services from IONOS or on-site servers. Currently, training still relies on standard CPUs and GPUs.