Ddos 
Red Hat has announced an expanded strategic partnership with AMD, uniting Red Hat’s leadership in open-source and AI software platforms with AMD’s advanced expertise in high-performance computing architectures. Together, they aim to deliver more agile and performance-optimized AI and virtualization solutions for enterprise clients, addressing the increasingly diverse demands of hybrid cloud and generative AI deployments.
In the realm of AI workloads, Red Hat affirmed that its OpenShift AI solution now offers full support for AMD Instinct GPUs. This enables enterprises to harness the power of high-performance GPUs without substantial resource investment, facilitating rapid deployment and implementation of generative AI models across hybrid cloud infrastructures.
The collaboration has also yielded successful demonstrations on Microsoft Azure’s ND MI300X v5 platform, equipped with AMD Instinct MI300X GPUs. These demonstrations showcased the inference capabilities of both small and large language models based on Red Hat Enterprise Linux AI. The ability to deploy multiple GPUs on a single virtual machine significantly reduces the complexity and cost associated with multi-VM setups.
Regarding vLLM—an open-source framework for efficient language model inference—Red Hat and AMD are deepening their involvement in upstream community collaboration. By optimizing the Triton core, incorporating FP8 support, and upstreaming core libraries, the partnership seeks to elevate inference performance and enhance collective communication and workload scheduling across multi-GPU environments.
Red Hat is also set to release its enterprise-grade vLLM distribution, the “Red Hat AI Inference Server,” which will be tightly integrated with AMD Instinct GPUs. This offering will empower enterprises to deploy open-source AI models on validated platforms with both stability and high performance.
On the virtualization front, Red Hat OpenShift Virtualization has now been certified on AMD EPYC processor platforms. This enables seamless integration of traditional virtual machine workloads into cloud-native architectures, allowing enterprise IT teams to modernize infrastructure within familiar environments while optimizing application deployment and resource allocation efficiency.
Looking ahead, Red Hat emphasized its commitment to deepening technological collaboration with AMD. The partnership will not only foster robust integration of AI model training and inference environments but will also accelerate enterprises’ transition to an AI-driven hybrid cloud era—realizing more flexible and cost-effective digital transformation strategies.
Donate