Intel Data Center GPU Flex 170 Single-Width PCIe Thermal

The Intel Data Center GPU Flex 170 Single-Width PCIe Thermal (GDCFLX170S) is built on Intel's Xe-HP multi-tile GPU architecture, delivering high-throughput media, analytics, and inference performance within a compact single-slot, single-width form factor designed for dense server deployments. Based on two Xe-HP GPU tiles interconnected via Intel's high-bandwidth tile interconnect, the Flex 170 provides substantial compute and media transcode throughput without the physical footprint of dual-width solutions, making it well-suited for space-constrained rack environments where GPU density is a priority.

The GPU integrates dedicated hardware media engines capable of simultaneous multi-stream AVC, HEVC, AV1, and VP9 encode and decode, positioning it as a purpose-built accelerator for cloud gaming streaming, video-on-demand transcoding, and broadcast media infrastructure. Its HBM2e memory subsystem delivers high bandwidth at relatively low power draw, supporting workloads that require fast data movement across large media or inference batches. The PCIe interface ensures broad platform compatibility across standard enterprise server form factors.

For enterprise IT, datacenter operators, and AI infrastructure teams across UAE, GCC, EMEA, and APAC regions, the Flex 170 Single-Width Thermal variant offers a thermally passive solution reliant on chassis airflow, simplifying integration into standard 1U and 2U servers with adequate airflow design. Intel's open software ecosystem, including the oneAPI toolkit and Intel Media SDK, enables workload portability and developer flexibility across heterogeneous compute environments, reducing dependency on proprietary programming models.

Ideal for

• High-density cloud gaming streaming with simultaneous multi-user AV1 and HEVC encode acceleration across dense GPU server configurations
• Video-on-demand and OTT transcoding pipelines requiring parallel multi-codec encode and decode at scale within limited rack space
• AI inference serving for vision and natural language processing models in latency-sensitive enterprise and edge datacenter deployments
• Broadcast and media production workflows demanding real-time multi-stream video processing with deterministic throughput
• Telco and communications infrastructure workloads such as 5G media gateway transcoding and remote desktop streaming services
• Enterprise virtual desktop infrastructure (VDI) acceleration providing GPU-backed remote graphics to large concurrent user populations

Technical specifications

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