AI PC New Era: Microsoft Copilot will be able to run locally on every computer
This technological advancement is not only an innovation in hardware, but also heralds a future where the depth and breadth of the integration of personal computer applications with AI functions will far exceed the current state, ushering in a new chapter of intelligent personal computing
Intel will help Microsoft break through the computational bottleneck of edge AI.
Intel recently revealed to the tech media Tom's Hardware that future AI PCs will need to be equipped with powerful built-in Neural Processing Units (NPUs) with processing capabilities of trillions of operations per second (TOPS) to support Microsoft's Copilot AI service for local execution on personal computers. This transformation aims to significantly reduce response latency, enhance performance, and strengthen user privacy protection.
This technological advancement is not only an innovation in hardware, but also indicates that the depth and breadth of the integration of AI functions with personal computer applications in the future will far exceed the current state, ushering in a new era of intelligent personal computing.
Currently, Copilot relies heavily on cloud computing, leading to delays even for small-scale requests. By enhancing local computing capabilities, future AI PCs are expected to provide a smoother user experience. In terms of hardware specifications, in addition to at least 16GB of memory, Microsoft is also expected to require the next generation AI PC to have at least 40 TOPS of processing power.
Although the current NPU applications in the Windows system are relatively limited, compared to ChromeOS and macOS, the latter two have widely used NPUs in advanced functions such as video and audio processing, including OCR, translation, and real-time transcription. In terms of processor performance, Apple's M3 series leads with a maximum NPU speed of 18 TOPS, while AMD and Intel's respective products have also demonstrated good performance.
Intel also revealed that the Lunar Lake chip planned for release in 2025 will triple the NPU speed, and announced 300 AI function optimizations for its OpenVino platform. In addition, the company has launched an AI PC development kit based on the ASUS NUC Pro, showcasing Intel's ambition and technical expertise in the field of AI PCs