Sugon announced at the 2026 Intelligent Computing Applications Conference that
it has commissioned Shuguang 8000 (Dengfeng), China’s first fully domestically
produced 100,000‑AI‑accelerator‑card supercluster, and has launched development
of a second 100k‑card system. The Shuguang 8000 adopts a “super‑intelligent
fusion” architecture to meet full‑precision, large‑scale compute demands for
large models and scientific AI, supporting FP64–INT8 precision. It targets
scientific computing, large‑model training, AI inference and industrial
simulation. The system is connected to the national integrated supercomputing
network; on its 100k‑card core nodes Sugon says it has completed 300+
application optimizations across 20+ frontier areas including large models,
robotics, quantum computing and new materials. More than 70 applications have
been scaled to ~10,000‑card deployments, validating core‑node stability and
reliability under large‑scale, high‑load research tasks.
