Microsoft opens preorders for Surface Laptop Ultra and RTX Spark Dev Box

Microsoft is framing both machines as expressions of its hybrid intelligence vision, in which a local PC and cloud computing resources are combined rather than relying on the cloud alone. The Surface Laptop Ultra is aimed at creators and AI engineers. It uses an NVIDIA RTX Spark superchip that consolidates an NVIDIA Blackwell RTX GPU with up to 6,144 cores, an NVIDIA Grace CPU with up to 20 cores, and up to 128GB of unified memory onto a single chip. Microsoft says this delivers up to 1 petaflop (FP4) of local AI processing performance, enough to run large AI models with more than 120 billion (120B) parameters directly on the device. That lets users perform local inference and evaluation using the CUDA ecosystem and the Windows AI stack while reducing latency and token costs tied to cloud communication.

The notebook is under 18mm thick and weighs about 2.04kg or less, and it carries a cooling system with up to 2.5 times the heat capacity of the previous generation. Its 15-inch PixelSense Ultra touchscreen has peak brightness of up to 2,000 nits. Ports include the Magnetic Connect magnetic detachable USB-C charging standard, plus HDMI, USB-A, and an SD card reader, with support for up to three 4K display outputs. The Surface RTX Spark Dev Box, a small desktop for AI engineers and technical teams, uses the same NVIDIA RTX Spark and 128GB unified memory. It is positioned as part of the Project Zenith developer initiative and ships with environments such as Visual Studio Code, Git, GitHub CLI, GitHub Copilot, WSL, Python, and Node already set up. To keep locally run AI agents safe, Microsoft provides isolation through Microsoft Execution Containers, and says it will later link with Microsoft Entra and Microsoft Agent 365 to manage user permissions and agent permissions separately.