Intel Data Center GPU Flex 170

The Intel Data Center GPU Flex 170 is a purpose-built discrete GPU based on Intel's Xe-HP microarchitecture, designed specifically for media delivery, cloud gaming, and AI inference workloads in enterprise data center environments. Manufactured on a 6nm process node, it integrates two Xe-HP compute tiles on a single card, delivering scalable parallel compute and hardware-accelerated media transcode capabilities suited to high-density rack deployments.

The Flex 170 features 32 Xe-cores across its dual-tile configuration, 24 GB of GDDR6 memory with an aggregate bandwidth of 576 GB/s, and hardware ray tracing support. The card supports AV1, HEVC, AVC, and VP9 encode and decode natively in silicon, enabling high-density video transcoding streams per card without CPU offload — a critical requirement for media service providers and cloud gaming infrastructure operators running at scale.

With a 150W TDP in a standard PCIe Gen 4 x16 full-height full-length form factor, the Intel Data Center GPU Flex 170 is engineered for server integration across major data center platforms. It is supported by Intel's open software stack including the oneAPI toolkit, OpenCL, and SYCL, providing enterprise AI and media teams a vendor-open programming model. The GPU is identified under manufacturer part number GPUFLEX170SSDLQCRX and is available through Omnixon Global for enterprise procurement across UAE, GCC, EMEA, and APAC regions.

Ideal for

• High-density video transcoding for OTT and media streaming platforms requiring hardware AV1, HEVC, and AVC encode/decode at scale
• Cloud gaming infrastructure deployments demanding concurrent GPU-accelerated game streaming sessions per server node
• AI inference at the edge or in the data center for computer vision, recommendation, and natural language processing models using oneAPI or OpenVINO
• Virtual desktop infrastructure (VDI) and cloud workstation environments requiring GPU-accelerated remote visualization
• Telecommunications and 5G vRAN workloads where GPU compute offload reduces CPU overhead in virtualized network functions
• Enterprise media production pipelines utilizing GPU-accelerated encoding for live broadcast, archiving, and adaptive bitrate ladder generation

Technical specifications

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