Microsoft is reshaping Windows around the rise of agentic AI, seeking to transform the operating system from a platform for running applications into an environment where AI agents can execute tasks on behalf of users. At an October 7 event in San Francisco, the company introduced new Windows, Copilot and Surface capabilities as part of a broader effort to move some AI workloads from the cloud directly onto personal computers.

Windows Moves Toward “Hybrid Intelligence”

Microsoft CEO Satya Nadella described the company’s new direction as “Hybrid Intelligence,” combining cloud-based AI capabilities with models and tools that operate directly on a PC. The approach is designed to bring together the model, software and device so AI can become more useful in everyday work.

Local processing could reduce reliance on cloud computing for tasks that do not require substantial remote computing power. Copilot, for example, could independently locate files needed to prepare tax returns, collect them from a computer and perform additional tasks without sending the information to the cloud. Microsoft argues that local processing can be faster and less expensive while potentially improving data privacy.

Copilot Takes a Larger Role Inside Windows

Microsoft is expanding Copilot’s integration into Windows and positioning it as a central interface through which users can interact with AI agents, launch AI tools, write code and complete tasks. A new capability called Autopilot allows users to give an agent a role and objective before allowing it to carry out tasks such as preparing project updates, tracking decisions and deadlines, reminding colleagues and gathering preparation materials.

The company is also adding security and control mechanisms intended to limit AI agents’ access to files and computing resources. Microsoft Execution Containers, designed to isolate agents and control their permissions, are now generally available in Windows. The move highlights one of the central challenges of agentic AI: as systems gain greater autonomy, controlling which actions they can take becomes increasingly important.

Surface Hardware Puts AI Computing at the Center

The strategy extends beyond software. Microsoft introduced the Surface Laptop Ultra, a premium computer designed for developers and creators that incorporates Nvidia’s RTX Spark technology. The device starts at $2,599 and will be available from October 16. Microsoft also unveiled the Surface RTX Spark Dev Box, priced at $5,999 and expected to launch in November.

The combination of Windows, Copilot, specialized hardware and Nvidia AI technology reflects Microsoft’s effort to capture a larger portion of the AI value chain. The company said in July that Microsoft 365 Copilot had surpassed 30 million paid seats, while net-new seat additions doubled quarter over quarter. Microsoft’s cloud business is also benefiting from AI demand, but the next test will be whether users and businesses adopt local AI capabilities at scale.

Going forward, investors will watch whether local AI processing and the ability to run agents directly through Windows create a meaningful commercial advantage for Microsoft. High hardware prices, the need for strong agent security and competition from Apple, Google, OpenAI and Anthropic will remain important factors. For Microsoft, successful adoption could strengthen its position not only in operating systems, but also across the software and hardware infrastructure underpinning the next phase of AI.