Sharp partners with Foxconn to enter the AI server market with launches planned for 2027. The enterprise model features a modular design supporting up to eight NVIDIA RTX PRO 6000 Blackwell GPUs, dual Intel Xeon 6 processors, and 3TB memory. The training model utilizes NVIDIA HGX Rubin NVL8 with liquid cooling and NVLink to integrate eight GPUs, supporting ConnectX-9 SuperNIC and BlueField-4 DPU configurations. Target customers include enterprises, the Japanese government, research institutions, and data centers.