In a recent interview, Microsoft CEO Satya Nadella laid out a bold vision for the future of personal computing: reinventing Windows entirely around AI “agents.” Speaking about the launch of the new Nvidia-powered Surface Laptop Ultra, Nadella detailed how these autonomous AI assistants will fundamentally change how we interact with our computers—and how Microsoft plans to keep them from going rogue.

Is AI Going to Kill Us?

When asked the blunt question of whether AI is going to kill us, Nadella was straightforward: “I think we have to do one simple thing, which is ensure that that never happens.”

He emphasized the concept of “pacing” in the tech industry—the idea that when building complex, potentially dangerous systems, companies must take their time. For Nadella, the tech industry’s “social permission to operate” relies entirely on its ability to build tools that serve humanity while remaining strictly under human control.

Agents as “First-Class Citizens” in Windows

The most radical shift coming to Windows is how the operating system treats AI. Rather than just being an app you open, AI agents will have their own identity on your computer.

Nadella explained that an AI agent needs to be treated almost like a separate user. “They need to be separate. They need to be contained separately. They need to be governed separately, permissioned separately,” he noted [04:48].

This allows users to track exactly what an agent is doing, what CPU power it is consuming, and what files it is accessing. By leveraging powerful local Nvidia chips alongside cloud computing—what Nadella calls a “hybrid moment”—these agents can work quickly and securely on your local device without always needing the internet.

The Need for “AI Antivirus”

With recent headlines about AI agents going rogue or hijacking websites, trust is a major hurdle. Nadella revealed that Microsoft is treating all AI models and agents as an “insider risk.”

To combat this, Microsoft is architecting secure containers (MXC containers) so that whatever happens with an agent stays isolated. Furthermore, Microsoft Defender is evolving into an AI antivirus. It will continuously monitor an agent’s behavior in the background. If the agent tries to perform an unauthorized action, Defender will instantly flag and shut it down [16:28].

Nadella’s Own AI Assistant: “SN Mo”

Nadella isn’t just preaching this future; he’s actively living it. He shared that he used a local AI model to code his own custom “dream to-do app.”

Even though he hadn’t written a WinForms app since the early 2000s, he prompted the AI to build a tracker that perfectly integrates his Outlook emails, Teams messages, and CRM data into one seamless dashboard. And as for his personal Copilot? He doesn’t call it Copilot—he named it “SN Mo.”

Beyond the Laptop: Project Solara

While the Surface Laptop Ultra represents the current powerhouse for developers and power users, Nadella teased that the form factors of computers are going to change. He highlighted an early prototype from Project Solara—a wearable computing badge with a fingerprint sensor.

Designed for frontline workers like nurses, this badge allows them to tap to unlock their AI agents and instantly streamline their workflows, such as updating electronic medical records on the go, without dragging a laptop around.

Ultimately, Nadella remains highly optimistic. When asked to close his eyes and picture the future of computers, he described a machine that simply does all the things we’ve always wanted to do but never got around to. And as for surviving the AI revolution? “I am alive because I made the computer do things that made this world a better place for all of us,” he concluded [23:50].