This is the way: Former Windows boss backs Apple's on-device AI push

Former Microsoft Windows chief Steven Sinofsky has indirectly praised Apple's AI strategy. Soon after the Cupertino major announced its Siri AI update powered by the next generation of Apple Intelligence leaning heavily on on-device AI, Sinofsky took to X to proclaim, "On device models FTW."

In Short

• Ex-Windows chief Steven Sinofsky says on-device AI FTW
• Sinofsky said on-device AI is the way most use-case is headed
• The comments came soon after Apple announced Siri AI

Apple concluded this year's developers conference, WWDC 2026, with a keynote packed with announcements, from iOS 27 and macOS Golden Gate to the long-awaited Siri AI and the next generation of Apple Intelligence. One of the biggest talking points from the event was Apple's focus on on-device AI. In fact, it was this approach that even caught the attention of former Microsoft Windows chief Steven Sinofsky, who praised on-device AI as the direction in which most AI use cases are headed.

Shortly after Apple's keynote, Sinofsky posted on X: "On device models FTW" (for the win), adding that he had "always been convinced this is where the 90% use case is heading."

His comments came after Apple detailed the next generation of Apple Intelligence, which places a strong emphasis on running AI tasks directly on users' devices rather than relying entirely on cloud-based systems.

What is Apple’s AI strategy?

At WWDC 2026, Apple introduced Siri AI, a completely rebuilt version of Siri powered by next-generation Apple Intelligence. The company explained that its new AI architecture combines powerful on-device models with Private Cloud Compute, allowing many tasks to be processed directly on iPhones, iPads and Macs.

Apple says this approach enables users to access AI-powered features while keeping personal data private. The company repeatedly highlighted that information processed on-device never leaves the user's hardware. When more computing power is needed, requests can be handled through Private Cloud Compute, which Apple says does not store user data.

The company also revealed that it had collaborated with Google and leveraged technologies behind the Gemini family of models to build the next generation of Apple Foundation Models. These models power features such as Siri AI, image generation, visual intelligence and system-wide writing tools.

Why on-device AI matters?

Much of today's generative AI ecosystem relies on cloud computing. Services such as ChatGPT, Gemini and Claude typically send user requests to remote data centres where large AI models process information and generate responses.

Apple, however, is taking a slightly different approach with Apple Intelligence. The company is betting that many everyday AI tasks can be performed locally on modern devices powered by Apple Silicon.

This means features such as personal context understanding, smarter search, advanced dictation, visual intelligence and parts of Siri AI can work without constantly sending information to external servers. Apple argues that this not only improves privacy but can also reduce latency and make AI features feel more responsive.

The company even offered a glimpse of how this will work during the keynote. Siri AI can search through messages, emails, notes and photos stored on a user's device to find information, complete tasks and provide personalised responses. Apple also demonstrated visual intelligence capabilities that allow Siri to understand what users are looking at through their camera or on-screen content.

- Ends