At AICC 2026, Inspur launched the Yuanmao SD200 Ultra supernode AI server, built on domestic AI chips, capable of running the 2.8-trillion-parameter Kimi K3 model with token generation latency breaking 5.85 ms for the first time. It also released the Yuanmao HC2000 multi-computing-power unit.
On the evening of Sept. 11, Inspur Information disclosed its 2026 private placement plan, proposing to issue up to 96 million shares to no more than 35 investors and raise up to RMB 9 billion for projects including next-generation AI infrastructure R&D and industrialization, liquid-cooled native computing infrastructure R&D, multi-agent collaborative storage-computing converged infrastructure R&D, and working capital replenishment.