Intel unveiled multiple chip-to-system innovations today at Computex 2026. The company introduced its new rackscale AI infrastructure to help customers expand inference workloads. This new rackscale AI infrastructure utilizes Intel Xeon processors. It also integrates SambaNova SN-50 Reconfigurable Dataflow Units to boost performance.
The shift toward agentic workloads alters hardware needs inside modern data centers. AI application production creates a massive demand for power-efficient setups. Consequently, central processing units are gaining renewed importance.
Creative Strategies CEO and principal analyst Ben Bajarin shed light on these shifting computing dynamics.
“While the training-era world looked closer to a one-CPU-per-four-GPU relation in AI deployments, agentic inference changes that relationship to roughly a one-CPU-to-one-GPU (or less) ratio.”
Intel, SambaNova, and Foxconn plan to build these hardware solutions together. This specific rackscale AI infrastructure targets hyperscale and intelligence center deployments. Foxconn will handle the system integration capabilities. Furthermore, Foxconn plans to manufacture a CPU-dense variant for data processing.
Expanding Strategic Industry Collaborations
Intel also highlighted partnerships with companies like Siemens and Hitachi. Siemens will expand its work with Intel into chip design and fab digitalization. Hitachi intends to collaborate on foundry tools. Meanwhile, Greenstone Biosciences will deploy Intel processors to accelerate drug development.
New chip innovations drew significant interest during the event. Intel announced the availability of its Xeon 6+ processors. This represents the first use of Intel 18A technology for a data center CPU. A single liquid-cooled rack offers 36,864 cores.
Intel CEO Lip-Bu Tan shared his perspective on the industry transformation.
“For more than five decades, Intel, its ecosystem partners, and Taiwan have brought the world the foundational technologies for the PC, Internet, and now AI eras. Today, with the rise of inference, agentic, and physical AI, Intel is poised to bring the world new innovations from the chip to systems level that promise to transform industry and society for the better. We are proud to join all our partners in building great products that will delight customers and bring the power of AI to more people as we create a brighter future together.”
Additionally, Vector Core Compute unveiled a new enterprise inference cloud. This cloud ran a real-world demonstration of fully disaggregated inference. The setup used Intel Xeon 6 processors, SambaNova SN40 RDUs, and NVIDIA Blackwell GPUs. Together.ai is the first commercial customer using this cloud. Finally, the platform achieved the fastest inference on the MiniMax 2.5 model.
Explore IT Tech News for the latest advancements in Information Technology & insightful updates from industry experts!
News Source: Businesswire.com
