The personal computing industry is aggressively shifting away from cloud-dependent software in favor of local, autonomous AI agents. Microsoft unveiled the Surface Laptop Ultra today at Computex in Taipei to meet this hardware demand. The 15-inch device weighs under 4.5 pounds and relies entirely on NVIDIA’s newly announced RTX Spark system-on-a-chip.
Microsoft abandoned the experimental hardware designs of its previous flagship machines. The detachable screens of the Surface Book and the complex hinges of the Surface Laptop Studio are gone. The Ultra uses a traditional clamshell chassis built specifically for high-end developers and creative professionals, according to a detailed hardware preview by Engadget. Microsoft Corporate Vice President of Surface Andrew Hill said, “This is the most powerful thing we’ve ever made.”
The machine includes a 15-inch MiniLED Ultra display capable of reaching 2,000 nits of peak HDR brightness. It features the largest trackpad ever built into a Surface product. Microsoft also included a comprehensive port layout, offering USB-A, USB-C, HDMI, and a full-sized SD card reader without requiring external dongles.
Inside the RTX Spark Architecture
The internal hardware marks a significant departure from standard Intel or AMD x86 processors. The NVIDIA RTX Spark combines 6,144 Blackwell GPU cores and 20 Arm CPU cores into a single piece of silicon. NVIDIA CEO Jensen Huang stated that PCs are reinventing themselves from manual input machines into devices designed to run autonomous AI agents natively. This is supported by the RTX Spark’s capability to deliver 1 petaflop of AI computing power.
The chip provides graphics performance comparable to a discrete RTX 5070 laptop GPU. It also offers extreme power efficiency, scaling its power draw from single digits during light tasks up to 80W under heavy workloads.
Memory Allocation and OpenShell Security
To run intensive AI software locally, the Surface Laptop Ultra configures up to 128GB of unified memory. This architecture allows RAM to be allocated freely across the CPU and GPU as needed. Developers can run 120-billion parameter AI models directly on the laptop hardware, as outlined in Microsoft’s official device announcement.
The laptop strictly adheres to Microsoft’s Copilot+ initiative requirements, which mandate a baseline 40 TOPS neural processing unit. It also utilizes NVIDIA’s newly announced OpenShell security primitives to safely isolate and execute AI agents on the primary device.
How the Surface Laptop Ultra Alters the Windows on Arm Race
This release changes the competitive structure of the premium laptop market. The Surface Laptop Ultra places NVIDIA in direct competition with Qualcomm’s Snapdragon X2 and AMD’s Ryzen AI Max within the Windows on Arm ecosystem. More directly, the Ultra targets Apple’s high-end MacBook Pro market share.
Historically, Microsoft’s top-tier hardware relied on unique physical form factors to differentiate itself from Apple. By stepping away from experimental hinges and focusing entirely on raw silicon performance with NVIDIA, Microsoft is competing purely on processing capability and unified memory limits.