Windows 11 Runs on Apple’s New iPhone Duo as Developer Bypasses iOS Limitations

Apple's first foldable iPhone has become an unlikely platform for Microsoft's desktop operating system, with a Google product lead demonstrating how Windows 11 can run locally on the device without jailbreaking it.

The project offers an early look at the foldable's hardware possibilities while exposing the software restrictions developers must navigate to run operating systems outside Apple's ecosystem.

The project, called WindowsiOS, was developed by Ammaar Reshi, who leads product and design at Google AI Studio. Reshi published a demonstration on X on November 11 and made the source code available on GitHub.

The demonstration shows a retail version of Windows 11 24H2 running Microsoft Edge and Google's Antigravity coding application on the iPhone Duo. The interface adapts to the phone's folding design. When the device is partially folded, the lower display functions as a keyboard and trackpad. Opened fully, Windows occupies the entire 7.6-inch screen, with applications controlled through touch.

The achievement is notable because Windows has previously run on iPhones only through slower emulation methods. In 2024, the developer behind Tiny11 managed to boot a stripped-down version of Windows on an iPhone 15 Pro using the UTM SE emulator. Reaching the desktop took about 20 minutes, largely because the software had to interpret Windows instructions rather than execute them directly through the phone's processor.

WindowsiOS takes a different approach, using an instruction translator to make Windows software run more efficiently on the iPhone's ARM-based hardware.

The Technical Workaround Behind WindowsiOS

Apple's restrictions on low-level system access have made running alternative operating systems on standard iPhones difficult. Hardware virtualization would normally provide a faster route, but Apple removed hypervisor access on stock iOS beginning with version 16.4.

Reshi's project instead uses a translator that converts Windows' Arm instructions into Arm instructions the iPhone Duo's A20 Pro processor can execute. The shared Arm architecture provides a foundation for the process, allowing the software to avoid the heavy instruction-by-instruction interpretation associated with traditional emulation.

The approach builds on Henge, an open-source project that has previously enabled macOS to run on modern iPhones. WindowsiOS adapts that existing work to support Microsoft's operating system.

The project's documentation credits Gemini 4 Carbon with handling the Windows-specific engineering. Google opened the model to developers and AI Ultra subscribers during the same week Reshi published the demonstration, making the project an early example of how AI coding tools can contribute to complex software development.

However, the underlying technology is not entirely new. Henge provides the core translation framework, while the Windows-specific implementation adapts it to another operating system. The demonstration therefore combines existing open-source engineering with newer AI-assisted development tools.

The result also has limitations. Windows receives only 4.5GB of the iPhone Duo's 12GB of memory. To fit within that allocation, Reshi removed more than 35 Windows background services, including Microsoft Defender. The changes reduce idle memory consumption to approximately 1.46GB.

That optimization is deemed crucial because running a desktop operating system on a mobile device creates competing demands for memory, processing power, and battery life. Removing background services makes the installation more practical, but it also means the Windows environment does not offer the same configuration or security features as a standard installation.

Installation Difficult, Performance Unverified

WindowsiOS is not a consumer application that can be downloaded from Apple's App Store. Installing it requires an Apple Silicon Mac running Xcode and a separate Windows 11 arm image supplied by the user.

Each launch also requires a debugger to remain connected to the iPhone. The connection allows the translator to access executable memory through a workaround that bypasses restrictions affecting applications that rely on just-in-time compilation. That requirement creates a potential point of failure. Apple has previously restricted access to system-level capabilities, including hypervisor support, and could close the route used by WindowsiOS through a future iOS update.

The project's performance claims also remain difficult to assess. Its README describes execution as approaching native speeds, but provides no benchmark results comparing WindowsiOS with conventional emulation or other operating environments. The software has only been tested on the iPhone Duo, leaving questions about compatibility with other iPhone models and future versions of iOS.

The absence of independent performance tests means the demonstration should be viewed as a technical proof of concept rather than evidence that the iPhone can replace a Windows computer. Running a desktop interface is only one part of that challenge. Sustained performance, application compatibility, battery consumption, thermal management and security would determine whether the experience could become useful beyond experimentation.

The project is also part of a growing exchange between devices and operating systems that manufacturers normally keep separate. In late October, another developer demonstrated iOS 26 running on Microsoft's Surface Duo 2. That developer has since indicated plans to bring the iPhone Duo's iOS 27 build to the Microsoft device.

These experiments show how hardware similarities and open-source development can create opportunities to run software outside the environments intended by manufacturers. They do not, however, remove the restrictions imposed by device security systems or establish that the resulting installations are ready for everyday use.