ADATA LEGEND 970 PCIe Gen5 NVMe SSD 2TB

The ADATA LEGEND 970 (SLEG-970-2000GCI) is a PCIe Gen 5 x4 NVMe 1.4 solid-state drive engineered for professionals who demand the highest available bandwidth from a client-class M.2 form factor. Built around a 3D NAND flash architecture and paired with a high-performance Gen 5 controller, the drive delivers sequential read speeds up to 10,000 MB/s and sequential write speeds up to 10,000 MB/s, effectively doubling the ceiling of its Gen 4 predecessors and saturating the PCIe 5.0 x4 interface available on current-generation AMD Ryzen 7000 and Intel Core Ultra platforms.

The LEGEND 970 2TB variant (SLEG-970-2000GCI) integrates a substantial DRAM cache to sustain consistent random I/O performance under mixed enterprise and creative workloads. The drive ships in the standard M.2 2280 single-sided format, making it compatible with high-end workstation motherboards, mobile workstations equipped with PCIe 5.0 M.2 slots, and compact AI development systems. A bundled graphene and aluminum heatsink helps manage the elevated thermal output inherent to Gen 5 operation, keeping the controller and NAND within safe operating temperatures during sustained transfers.

Targeted at AI model training prep, large-dataset ingestion, real-time 8K media production, and software compilation pipelines, the LEGEND 970 sits at the intersection of workstation and near-edge infrastructure storage. Its NVMe 1.4 command set supports namespace management and supports the low-latency queue depth characteristics required by latency-sensitive development toolchains, making it a technically credible choice for organizations provisioning high-throughput local storage for data scientists, ML engineers, and content production teams.

Ideal for

• AI and machine learning development workstations requiring rapid dataset loading and checkpoint storage at sustained multi-GB/s throughput
• 8K and high-frame-rate video editing rigs where real-time playback of uncompressed or minimally compressed media demands consistent sequential read performance exceeding 8,000 MB/s
• Large-scale software build and compilation environments where fast random read I/O reduces iteration cycle times for engineering teams
• Data science pipelines performing high-speed ingestion and preprocessing of multi-terabyte structured and unstructured datasets on local NVMe scratch storage
• CAD, CAE, and simulation workstations loading large finite-element or CFD model files where storage latency directly impacts time-to-first-frame
• Edge inference nodes and compact AI appliances that require maximum M.2 storage bandwidth within a thermally constrained 2280 single-slot footprint

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.