Can one storage platform handle AI datasets, backups, media, and virtualisation? An Xeon-based NAS range is designed to consolidate these workloads.

TerraMaster has launched its 725 Series enterprise rackmount NAS range, combining Intel Xeon D processors, ECC memory, high-speed networking, NVMe storage, and PCIe expansion to support data-intensive enterprise workloads.
The four models are designed to consolidate file storage, backup, disaster recovery, media production, AI training datasets, surveillance, virtualisation, and enterprise applications on a single platform. This approach can reduce the need for separate storage systems while simplifying infrastructure management as data volumes and concurrent workloads increase.
The hardware is built around Intel Xeon D-series processors with 16GB of ECC memory across the range. The U8-725 and U12-725 use four-core Xeon processors, while the U12-725 Plus and U16-725 Plus use eight-core variants for workloads requiring greater concurrency.
Networking and storage options are also intended to support high-throughput applications. The U12-725, for example, includes four 10GbE SFP+ ports and two 5GbE ports, alongside support for PCIe 4.0 NVMe SSDs and PCIe expansion. Under TerraMaster’s laboratory testing conditions, the model reached sequential read speeds of up to 6,250 MB/s and sequential write speeds of up to 3,480 MB/s. Actual performance will vary with configuration and workload.
The range spans eight to 16 drive bays, allowing enterprises to select capacity and processing capabilities according to their storage requirements. The U12-725 and its Plus variant target higher-performance applications, while the U16-725 Plus is positioned for larger-scale enterprise data environments.
The systems run TerraMaster’s TOS 7 operating system, which includes data search, indexing, and management capabilities. The platform can also provide centralised storage for AI training datasets and local data used by private AI applications.
Click here for the official announcement.





