INTEL CORPINTC

Intel advanced agentic AI infrastructure with rack-scale solutions and demonstrated production-ready systems.

“▪ Intel advanced its agentic AI infrastructure strategy with new rack-scale AI infrastructure and disaggregated inference solutions built on Intel® Xeon® processors. Intel, SambaNova and Foxconn demonstrated production-ready rack-scale infrastructure for inference and agentic workloads, while Vector Core Compute (VC2) unveiled a disaggregated agentic cloud combining Intel® Xeon® processors, SambaNova RDUs and NVIDIA Blackwell GPUs.”

Intel expects rapid adoption of agentic AI driving demand for compute silicon including GPUs, ASICs and CPUs.

“We expect these trends to continue in 2026 and future years, with continuing rapid adoption of AI technologies across an increasing number of industries and applications driving demand across a full spectrum of AI workloads, including generative AI, AI inference, agentic AI and physical AI, from traditional and AI data centers to network and enterprise environments and client and edge. We believe AI represents a generational shift in computing requiring a range of compute silicon, including GPUs, ASICs, CPUs and xPUs. Our multi-year strategy is designed to enable us to participate in the expanding market and increase in compute demand through our products and our foundry capabilities, including wafer manufacturing and advanced packaging. Our strategy is centered on four core priorities:”

Intel advancing x86 portfolio for inference, agentic and physical AI and partnering with NVIDIA on custom products.

“Our x86 architecture and related CPU offerings have served as a foundational computing platform for over four decades, supporting a broad range of client and data center applications and enabling a heterogeneous environment from which a deep and extensive software ecosystem has developed. The widespread adoption of our x86 architecture and related CPU offerings and established software ecosystem continues to provide interoperability and performance across diverse workloads. We remain focused on adapting and advancing our x86 product portfolio to meet evolving market needs, including those driven by inference AI, agentic AI and physical AI. In 2025, we introduced new x86 solutions tailored to client and data center platforms, including our next generation client CPU code named Panther Lake, which is designed to combine improved power efficiency and performance to support next-generation AI-enabled workloads at the edge. We also announced a strategic partnership with NVIDIA to co-develop custom client and data center products combining our x86 CPU technologies with NVIDIA's AI and accelerated computing capabilities. The collaboration is intended to jointly develop multiple generations of products for hyperscale, enterprise and consumer markets based on the x86 architecture.”

Intel aims to develop ASICs, GPUs and design services for agentic AI workloads.

“We aim to leverage our engineering and design expertise to develop purpose-built ASICs and GPUs for customers to address the expanding variety of AI-driven compute workloads. In 2025, we unified our horizontal engineering functions into a central engineering group to drive efficiencies across foundational IP development, test chip design, EDA tools and design platforms to help us deliver the combination of infrastructure and software needed for agentic AI workloads, which require integration across multiple models and compute types. With this new internal consolidation of capabilities, we also aim to develop an ASICs and design services business to deliver purpose-built silicon for a broad range of external customers. The goal is to both extend the reach of our core x86 IP and leverage our design strengths to deliver an array of solutions from general purpose to workload-optimized computing.”

Intel aims to support emerging inference, agentic and physical AI workloads with its platforms.

“Inference AI, agentic AI and physical AI workloads are rapidly emerging areas that we expect may ultimately represent larger market opportunities than the generative AI workloads that have been the core area of AI technology growth over the last few years. We aim to partner with an array of incumbent and developing companies defining these emerging areas and position our x86 platform to be their platform of choice. We also aim to participate in these emerging areas through our continued development of our Xeon, AI PCs, Arc GPUs and our open software stack, as well as by developing successive generations of inference-optimized GPUs on a targeted annual cadence that feature enhanced memory and bandwidth to meet enterprise needs.”

Intel is developing heterogeneous xPUs to support inference, training and agentic AI models.

“We believe the future of computing lies in complex systems utilizing a diverse mix of scalar, vector, matrix and spatial architectures deployed in CPUs, GPUs, NPUs, IPUs and accelerators, which we refer to generally as xPUs, supported by a scalable software stack and integrated through advanced packaging technologies. This approach is designed to address AI, analytics and high-performance workloads. In support of this approach, we are developing processors that incorporate multiple computing architectures, moving toward heterogeneous computing where x86 CPUs operate alongside specialized AI engines to support performance for inference, training and emerging agentic AI models.”