Microsoft’s first major Windows event in over two years came down to one laptop and one number. The Surface Laptop Ultra, built around Nvidia’s new RTX Spark chip, starts at $2,599, and the number that matters most is 128: gigabytes of unified memory.
What is the Surface Laptop Ultra?
It is Microsoft’s most powerful Surface laptop, announced on October 7 in San Francisco with Nvidia CEO Jensen Huang on stage. Inside is the RTX Spark platform, which combines an Nvidia Grace CPU, with up to 20 cores, and a Blackwell RTX GPU. It supports up to 128GB of unified memory and, by Microsoft’s numbers, up to one petaflop of AI performance at FP4 precision.
The 15-inch mini-LED touchscreen has a 3:2 aspect ratio and reaches up to 2,000 nits of peak brightness. The laptop is under 18 millimeters thick and weighs under 4.5 pounds. Microsoft also says it retains 99% of its power on battery, though that claim awaits independent testing.
How much does it cost?
There are eight configurations, from $2,599 to $5,899. Preorders are open and shipping starts on October 16. A compact Surface RTX Spark Dev Box for developers costs $5,999.
The memory upgrade explains much of the spread: going from 64GB to 128GB costs $1,600, which works out to $25 per extra gigabyte. That is a useful illustration of what the memory shortage is doing to prices.
Laptops with the same chip from Lenovo, Asus, Dell, HP and MSI go on preorder today and ship on October 16.
What does Microsoft claim?
According to Microsoft, the laptop answers locally run AI prompts twice as fast as a comparable MacBook Pro M5 and generates video six times faster. Bloomberg reports it costs about $400 less than an entry-level 16-inch MacBook Pro. These are company claims, since no independent reviews exist yet.
The software side matters too. Microsoft announced Windows 11 changes for agentic AI, an overhaul of Copilot, a local coding model and safeguards for running AI agents safely on the PC.
Who is it for?
Not most people, as Engadget put it. The current Surface Laptop starts at $1,200, so this model is aimed at developers, creators and anyone who wants to run large AI models locally instead of in the cloud. Memory capacity is the real selling point: big models need a lot of it, which is why 128GB matters more here than the clock speed of the processor.
Questions remain. Nvidia’s Grace CPU is Arm-based, so reviewers will need to show how well existing Windows software runs. Battery life under heavy AI load and the noise of the cooling system are also unknown.
Why it matters
The laptop shows how the idea of an AI PC is changing. A year ago it meant a small neural chip for background tasks. Now Microsoft and Nvidia are selling a machine whose headline spec is the amount of memory it can hold, and they are going after Apple’s most expensive laptops directly.
It also shows where the pressure is. The component that decides what a PC can do with AI is the same one in short supply, which pushes up prices across the industry. The Surface Laptop Ultra is a high-end preview of that trade-off, and the real test is how it performs when the first reviews arrive.