Microsoft's New AI Laptops Cost $2,599 and Up: What You're Paying For

Microsoft and PC makers have now priced laptops built to run AI helpers on the device. The first machines with Nvidia's RTX Spark chip, led by the Surface Laptop Ultra alongside models from Dell, Lenovo, MSI, HP and Asus, start at $2,599 and arrive as early as October 16th.
That $2,599 starting model, detailed by The Verge, has an 18-core processor, 24GB of shared memory and 512GB of storage. Shared means the processor and graphics use the same memory pool. The price is above a normal ultrabook and in mobile workstation range. A 48GB Surface Laptop Ultra is listed at $3,999.99 on Best Buy.
At the top end shared so far, an Asus ProArt P16 with a 20-core chip, 128GB of shared memory and 2TB of storage costs $6,999.99.
There are two chip choices. Specs outlined by Windows Central list an 18-core version with 5,120 graphics cores and a 20-core version with 6,144 graphics cores for the Surface Laptop Ultra. Microsoft describes the system as using a new NVIDIA chip that combines an ultra-efficient CPU with a powerful RTX GPU, according to Microsoft. The Surface Laptop Ultra supports up to 128GB of shared memory.
Several makers are in from the start. Nvidia lists the Dell XPS 16, HP OmniBook Ultra 16, Lenovo Yoga Pro 9n and MSI Prestige N16 Flip AI among RTX Spark laptop models. Nvidia had previously said RTX Spark laptops and compact desktops will be available this fall from ASUS, Dell, HP, Lenovo and Microsoft. For software makers, Microsoft also offers a Surface RTX Spark Dev Box with 128GB of shared memory.
Nvidia says RTX Spark gives 1 petaflop of AI performance, a measure of AI math speed, for what it calls the world's first Windows PCs purpose-built for personal agents, as noted on the Nvidia blog. Microsoft scheduled a special Windows and Surface event on October 7 covering the Surface Laptop Ultra and RTX Spark.
The broader context here is that memory size is the real divide between models. The step from 24GB to 48GB and 128GB is not about normal multitasking. For on-device helpers, shared memory decides what can stay ready at once, including the AI model, its short-term notes, search files and background helpers. Like a bigger desk, more memory lets more work stay close instead of going to storage or the internet.
In my view, these prices limit first buyers to developers and technical workers who already pay workstation prices to keep private AI work on their own machines. That is a sensible starting point. On-device work brings faster response, offline use and data staying at home, which matters more for helpers that do tasks for you. If the software delivers helpers that keep running and use local tools well, the bigger memory models will look less like a luxury and more like the normal setup. Over time, local machines tend to get more able, and this tests what early buyers will pay.


