Intel Xeon 6 Processors Target GPU-Accelerated AI Systems
Intel has launched three new Xeon 6 processors as enterprises grapple with the computational demands of large language models and AI training workloads. The chips incorporate Priority Core Turbo technology, which Intel claims will address a persistent bottleneck in AI systems where CPUs struggle to keep pace with increasingly powerful GPUs.
As AI models grow larger – with some now exceeding 400 billion parameters – the traditional approach of pairing any CPU with high-end GPUs has proven inadequate. The result has been expensive GPU clusters sitting idle whilst CPUs process data: a problem that has prompted both hardware vendors and cloud providers to rethink system architecture.
The Xeon 6776P processor has already found its way into Nvidia’s latest DGX B300 systems, marking a notable shift for Nvidia which has previously shown less dependence on specific CPU partners for its flagship AI systems.
“These new Xeon SKUs demonstrate the unmatched performance of Intel Xeon 6, making it the ideal CPU for next-gen GPU-accelerated AI systems,” says Karin Eibschitz Segal, Corporate Vice President and Interim General Manager of the Data Center Group at Intel. “We’re thrilled to deepen our collaboration with Nvidia to deliver one of the industry’s highest-performing AI systems, helping accelerate AI adoption across industries.”
Priority Core Turbo addresses AI workload bottlenecks
The Priority Core Turbo feature represents Intel’s attempt to solve a fundamental mismatch in how AI workloads actually behave versus how traditional CPUs operate. Many AI tasks require bursts of intense sequential processing followed by periods of parallel computation – a pattern that has historically left CPU cores either overwhelmed or underutilised.
The three new processors offer up to 128 Performance-cores each, a specification that puts them amongst the highest core-count CPUs available for AI applications. More significant may be the memory performance improvements – Intel claims 30% faster speeds compared to AMD’s EPYC processors in equivalent configurations.
Intel says the processors also include 20% more PCIe lanes than previous Xeon generations, addressing another common complaint from AI developers who have struggled with data transfer bottlenecks when connecting multiple GPUs or high-speed storage arrays.
Intel Computex announcements expand GPU and AI accelerator portfolio
The Xeon 6 launch follows Intel’s broader push into AI hardware announced at Computex 2025. The company unveiled Arc Pro B60 and B50 graphics cards aimed at AI inference workloads: a market segment that has seen explosive growth as organisations deploy trained models at scale.
The Arc Pro B-Series cards, with 24GB and 16GB of memory configurations, target a different segment than Nvidia’s flagship GPUs. Rather than competing directly with the A100 or H100 series, Intel appears focused on inference applications where cost and power efficiency matter more than raw training performance.
At Computex, Intel also announced availability of Gaudi 3 AI accelerators in both PCIe and rack-scale configurations. The Gaudi line represents Intel’s most direct challenge to Nvidia’s dominance in AI training, though adoption has remained limited compared to more established alternatives.
The company also made Intel AI Assistant Builder available on GitHub, enabling developers to create AI agents optimised for Intel hardware. Such software initiatives reflect industry recognition that hardware alone cannot drive adoption – developers need tools that make it easier to deploy AI applications on specific platforms.
“The Intel Arc Pro B-Series showcases Intel’s commitment in GPU technology and ecosystem partnerships,” says Vivian Lien, VP and General Manager of Client Graphics at Intel. “With Xe2 architecture’s advanced capabilities and a growing software ecosystem, the new Arc Pro GPUs deliver accessibility and scalability to small and medium-sized businesses that have been looking for targeted solutions.”
Explore the latest edition of Data Centre Magazine and be part of the conversation at our global conference series, Tech & AI LIVE.
Discover all our upcoming events and secure your tickets today.
Subscribe to our Weekly Newsletter for the latest insights, trends and expert analysis.
Data Centre Magazine is a BizClik brand
