Sharp has announced its entry into the AI server market in collaboration with Foxconn. The upcoming hardware lineup features 2 enterprise servers equipped with NVIDIA RTX PRO 6000 Blackwell GPUs and liquid cooling architecture. Shipments are scheduled to begin in 2027 for commercial clients, research institutions, and government agencies.
Sharp teamed up with Foxconn to build a modular system for standard enterprise workloads. It supports up to 8 NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. Customers can adjust internal components depending on their specific compute needs.
Processing power relies on dual socket Intel Xeon 6 processors backed by 3TB of memory spread across 32 individual 96GB DDR5 RDIMM sticks. Storage capacity sits at 38.4 TB. To keep the hardware running under heavy loads, the chassis houses 4 power units rated at 3200W each.
A second server targets intensive AI model training and large scale inference. This build runs on the NVIDIA HGX Rubin NVL8 platform. It links 8 GPUs inside a single enclosure using high speed NVLink connections. Heat management is handled entirely through liquid cooling. For network throughput and data processing, the server supports NVIDIA ConnectX 9 SuperNIC and NVIDIA BlueField 4 DPU configurations.
Sharp is directing these hardware solutions at enterprise data centers, academic research organizations, and Japanese government offices. Commercial sales will start in 2027. By combining its brand reach with the manufacturing scale of Foxconn, Sharp plans to establish a direct presence in the growing compute infrastructure sector.
