-
XTT657V2 AI server
The XTT657V2 is a flagship general-purpose server designed for domestic information innovation, government, and confidential applications, based on the latest Phytium server processing chip (S5000C-64C). It supports two Phytium S5000C 64-core processors, built on the latest FTC862 core architecture, offering more than double the computational performance of the previous S2500 generation. With diverse storage configurations and robust GPU and I/O expansion capabilities, the TG657V2 is suitable for various needs such as AI large model training, cluster training, inference, HPC high-performance computing, and graphics rendering.
Send Email Details -
XTT657V2-L AI server
The XTT657V2-L server is a flagship AI liquid-cooled server developed based on Phytium’s latest S5000C-64C server processing chip and fully domestically produced hardware components. It is designed for applications such as large model training, inference, HPC, and rendering. Featuring two Phytium S5000C 64-core processors built on the latest FTC862 core architecture, the server delivers at least twice the computational performance compared to the previous S2500 generation. It supports 8 to 10 full-height, full-length GPU cards, making it well-suited for large-scale AI operations in computing centers, industry markets, and internet sectors.
Send Email Details -
XTT9236 AI server
The XTT9236 is a high-performance GPU server developed on Intel's latest Eagle Stream platform. It supports NVIDIA H20 modules, Intel Gaudi modules, and other GPUs designed for large model training. It is the optimal choice for AI training and large model inference applications.
Send Email Details -
XTT8230 AI server
The XTT8230 is a powerful and flexible GPU cloud server developed on Intel's latest Eagle Stream platform. It supports multiple CPU-GPU and GPU-GPU interconnect topologies, catering to a wide range of AI business application needs.
Send Email Details -
XTT6237 AI server
The XTT6237 is a 2U AI product developed on Intel's latest Eagle Stream platform, designed for modern data centers. It supports 4 dual-width GPUs or 6 single-width GPUs, catering to AI training and inference scenarios in front-access data center environments.
Send Email Details
