Intel showcased details of its upcoming Meteor Lake chips of PC processors during Microsoft’s Build 2023 conference. With a “chiplet” system-on-chip (SoC) design, Intel aims to deliver advanced intellectual properties (IPs) and segment-specific performance while maintaining lower power consumption. Meteor Lake will introduce Intel’s first PC platform with a built-in neural VPU, enabling efficient AI model execution.
The integrated VPU will collaborate with existing AI accelerators on the CPU and GPU, allowing for accessible and impactful AI features for PC users. Intel asserts that its product is at the forefront of the AI trend, positioning Meteor Lake as a key player.
At Computex, Intel disclosed that the VPU in Meteor Lake is derived from Movidius’s third-generation Vision Processing Unit (VPU) design. By leveraging this acquisition from 2016, Intel aims to establish itself as an AI market leader. Although specific performance figures and VPU specifications have not been revealed, it is anticipated that Intel’s VPU will surpass Movidius’s previous throughput rating of 1 TOPS (tera operations per second).
As the VPU is integrated into the SoC, AI capabilities will be a standard feature across all Meteor Lake SKUs, rather than a differentiating factor. Intel seeks to achieve similar energy efficiency levels as smartphone SoCs, enabling tasks like dynamic noise suppression and background blur.
Collaborating closely with Microsoft, Intel aims to scale Meteor Lake and Windows 11 across the ecosystem. Through partnerships and leveraging the ONNX Runtime—an open-source library for deploying machine learning models—Intel plans to optimize AI model execution on the Windows platform.
Intel envisions shifting server-based AI workloads to client devices, offering benefits such as reduced costs, lower latency, and enhanced privacy. By pursuing this vision, Intel aims to gain a competitive advantage in the market.
