The right model at the right moment: Vendors push their agentic AI PC vision
AI is getting expensive, and companies are looking to cut token costs. The solution: move more AI work to PCs. The PC won’t replace the cloud. Preliminary AI work can be done on a PC, and tougher work — which today’s PCs can’t handle — can go to the cloud. “We’ll always need to use the most advanced cloud-based frontier models for certain tasks, but… we don’t need to use them for everything,”…
The growing demand for AI is driving companies to find ways to reduce token costs. The solution, according to many, is to move more AI work to personal computers (PCs). However, while PCs won't replace the cloud, they can handle preliminary AI tasks while more complex jobs can be sent to the cloud. Microsoft's vision of "hybrid intelligence" aims to make AI always available on Windows machines, with the cloud handling the hardest tasks and less demanding tasks running locally on PCs.
Microsoft refers to this as "agentic AI PCs," which can run agents, generate video, write code, and chat with advanced generative AI models on a PC. If the PC can't handle a task, it is sent to the cloud. These agentic AI PCs are equipped with powerful GPUs, similar to those in data centers. Nvidia's RTX Spark superchip, with a Blackwell GPU, powers many of these PCs, including Microsoft's new Surface Laptop Ultra.
Unlike older AI PCs, these new agentic PCs can run full AI models and have enough compute power to run them quickly. This gives users and enterprises a path to advance AI projects without incurring high cloud token bills. Companies like HP are also jumping on the bandwagon, with their new ZBook Ultra G3a mobile workstation featuring an AMD GPU that generates free tokens and cuts cloud bills.
The ZBook Ultra G3a can run agents that link local models to frontier models in the cloud, allowing for advanced AI-based design work without an internet connection.
Microsoft's CEO, Satya Nadella, emphasized that the future of AI is not about having just one model but having the right model at the right moment. This could be frontier capability when needed or smaller, faster models when speed is the priority. The goal is to seamlessly integrate AI into computing, making it available wherever the user needs it.
Written by urgent.news from Computerworld's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
