Intel Data Center GPU Flex 140 Dual Slot PCIe Thermal Solution

The Intel Data Center GPU Flex 140 Dual Slot PCIe Thermal Solution (GDCFLX140S) is a purpose-built enterprise graphics accelerator based on Intel's Xe-HPG microarchitecture, designed to deliver flexible, high-density media and cloud gaming workloads within modern data center environments. Built on a dual-slot PCIe form factor, it integrates dedicated hardware-accelerated AV1, HEVC, AVC, and VP9 encode and decode engines, enabling exceptional transcoding throughput per watt at scale. The Flex 140 is engineered for rack-dense deployments where thermal efficiency and workload versatility are critical operational requirements.

The GPU features 12 GB of LPDDR5 memory accessible via a 192-bit memory interface, delivering the bandwidth necessary for parallel media processing pipelines and virtualized desktop infrastructure at scale. With support for PCIe 4.0 x8 host connectivity and a single-card power envelope of 75W, the Flex 140 operates entirely within PCIe slot power without requiring supplemental power connectors, simplifying integration into standard server platforms. Intel's built-in open-source driver stack under Linux and integration with Intel oneAPI toolkits and OpenVINO frameworks further reduce adoption friction for enterprise software teams.

Targeted at cloud service providers, telecommunications operators, and enterprise IT teams running virtualized media pipelines, the Intel Data Center GPU Flex 140 supports simultaneous multi-user workloads through SR-IOV hardware virtualization, enabling multiple isolated virtual machine instances to share a single physical GPU without performance interference. Its compatibility with Intel's GPU pooling architecture and management tooling makes it well suited for large-scale, heterogeneous data center deployments across the UAE, GCC, EMEA, and APAC regions where media-rich application delivery and cloud gaming services demand consistent, low-latency GPU resources.

Ideal for

• High-density video transcoding at cloud scale using hardware-accelerated AV1, HEVC, AVC, and VP9 encode/decode engines for streaming media platforms
• Cloud gaming service delivery leveraging SR-IOV virtualization to provide isolated, low-latency GPU instances to multiple concurrent end users per card
• Virtual desktop infrastructure (VDI) deployments requiring shared GPU resources across large numbers of simultaneous remote enterprise users
• AI inference at the edge using Intel OpenVINO-optimized models for computer vision, object detection, and media analytics pipelines
• Telecommunications media gateway workloads demanding high encode throughput and energy-efficient operation within strict rack power budgets
• Broadcast and over-the-top (OTT) media processing workflows requiring parallel multi-format transcoding within dense, thermally constrained server chassis

Technical specifications

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