Recently, Annie Shea Weckesser, Chief Marketing and Communications Officer of Intel, delivered a keynote speech titled "Intelligence Ushers in the AI Era" at the 2026 Intel Technology Innovation and Industry Ecosystem Conference. In her speech, Annie Shea Weckesser introduced Intel's three-pronged strategy to collaborate with the industry ecosystem to drive the implementation of intelligence: first, building an open computing platform covering PCs, edge, and data centers; second, designing custom chips tailored for specific workloads; and third, leveraging process and packaging technologies to enable increasingly complex system integration in Intel products.
Intel is building an open platform covering PCs, edge, and data centers, enabling intelligence to run where it can best balance performance, power efficiency, security, and experience.
In personal computing. For most people, the PC is the earliest platform where intelligence integrates into daily work and creation. The 3rd Gen Intel Core Ultra processor provides the computing power required for complex reasoning on PCs, capable of running 35-billion-parameter models locally, while also featuring multimodal perception capabilities to understand context through visual, auditory, and other inputs. The 3rd Gen Intel Core Ultra processor integrates CPU, GPU, and NPU computing capabilities, allowing software to match each task with the most suitable compute engine, achieving a balance among user experience, performance, and power efficiency.
Moreover, developers, software providers, OEMs, and system manufacturers are joining forces to connect models, applications, and devices, enabling applications to run better on devices. In China, Intel is strengthening its collaboration with software partners such as DuMate, Marvis, Qianwen Office, TRAE, and WorkBuddy to deploy and optimize the new generation of hybrid agentic AI.
In physical AI and edge computing. Through physical AI and edge computing, intelligence enters the real-world scenarios of people's lives and work, from stores and factories to hospitals and cities. As intelligence becomes embodied, it can navigate environments, operate equipment, or assist in completing specific tasks.
The 3rd Gen Intel Core Ultra processor and Intel's open edge software help partners adapt intelligence to specific devices, sensors, and application environments. With powerful computing capabilities, fast response performance, and built-in inference acceleration, the processor supports demanding robotic workloads. Tools such as the Edge and Embodied AI Software Suite help partners optimize and deploy these workloads.
For example, leveraging the Intel Core Ultra processor, JAKA utilizes Kernel-based Virtual Machine (KVM) to integrate real-time motion control, perception, autonomous planning, and execution capabilities into a unified edge computing architecture. This enables robots to precisely control movements, understand their surroundings, and execute complex tasks across multiple scenarios.
In data centers and AI infrastructure. AI workloads continuously flow among planning, retrieval, reasoning, tool calling, and specific actions. Accelerators handle burst loads in model computation; the CPU coordinates various tasks as computation unfolds, driving concurrent workflows continuously toward final task completion. As AI workloads continue to scale, the system's demand for CPUs is also growing, and the GPU-to-CPU configuration ratio is evolving from 4:1 gradually towards nearly 1:1.
The CPU runs through the entire system, serving as the infrastructure that connects applications, data, orchestration capabilities, and dedicated accelerated computing resources. This is also a key reason why x86 will become even more important in the AI era. Intel estimates that by 2030, over 80% of servers globally will run on the x86 architecture.
The Intel Xeon 6 processor is built precisely to fulfill this role. The Performance-cores (P-cores) provide the speed needed for workloads requiring fast execution and response; the Efficient-cores (E-cores), built on the same secure and reliable Xeon architecture, support higher agent density and better power efficiency. Built-in accelerators handle routing at the host level, managing tasks such as security guardrails, compression, encryption, and confidential execution, enabling GPUs and other accelerators to focus on model computation.
The collaboration between Alibaba Cloud and Intel is a prime example. Alibaba Cloud is the first cloud provider in China to launch public cloud services based on the Intel Xeon 6 processor. Currently, Alibaba Cloud has launched over 117 cloud instance types based on Xeon 6, covering multiple product families including general-purpose computing, local SSD, high clock frequency, and confidential computing. Among them, more than 50 instance types are already widely used in agentic AI scenarios, helping customers drive agents from the lab to production deployment, providing a scalable and trustworthy computing foundation.
The second part of Intel's strategy is to design custom chips for customers and partners, including hyperscale cloud service providers, based on the platform computing foundation.
Intel is continuously investing in architecture, design, IP, Advanced Packaging, and manufacturing capabilities to create customized solutions for specialized deployments and intensive workloads. Here, Intel extends an invitation to the industry: "If you are developing differentiated solutions beyond standard platforms, we look forward to exploring together with you."
The third part of Intel's strategy is the process and packaging technologies behind Intel products.
AI is driving higher demands for greater computing power and better power efficiency across various products.
In chip manufacturing, process technology remains key to meeting this demand. Intel 18A has laid the technological foundation for next-generation Intel products and has entered high-volume manufacturing. Its high-performance version, Intel 18A-P, has also entered risk production. Intel 14A is also progressing smoothly. Risk production for internal products is planned to start in the second half of 2027, with the goal of achieving high-volume manufacturing ramp-up in 2028.
Advanced Packaging makes it possible to integrate the heterogeneous products required for AI workloads. Technologies such as EMIB and EMIB-T enable Intel to integrate multiple compute engines and IP into a single package, achieving efficient flow of data and workloads across the system through high-speed interconnects.
Intel is partnering with over 70 independent software vendors to drive the development of the AI PC experience, including 17 day-zero adaptation releases achieved jointly with Chinese AI model providers. Chinese cloud service providers have launched over 200 Xeon 6 and Xeon 6+ cloud instances. At the edge, Intel's collaboration with more than 100 robotics partners is driving intelligence into real-world environments.
Author | Ji Xiaoting
Editor | Zhang Xinyi
Art Editor | Ma Liya Supervisor | Lian Xiaodong