Microsoft opened pre-orders for the Surface Laptop Ultra on 7 October 2026, starting at $2,599 (MSRP), with availability beginning October 16, it said on its Windows blog. It also opened pre-orders for the Surface RTX Spark Dev Box at $5,999, exclusively on Microsoft.com in the US, shipping in November.
Both are built around NVIDIA’s RTX Spark Superchip. Microsoft says each can run AI models exceeding 120B parameters locally with up to 1 petaflop of AI performance and up to 128 GB of unified memory.
Business Pill · ON-DEVICE AI
A one-minute explainer of on-device AI: running a model on the phone or PC itself instead of in a data center, and what that changes for cost, speed and privacy. It teaches the general idea only and says nothing about any company in this story.
The key insight: As we read it, Microsoft is selling the laptop as a place to run tokens, not only as a laptop. Its Windows post says customers’ needs are outpacing their cloud budgets, and presents local models on RTX Spark as a way to make tokens go further.
The Surface Laptop Ultra pairs an NVIDIA Blackwell RTX GPU with up to 6,144 cores and an NVIDIA Grace CPU with up to 20 cores, according to Microsoft. Its footnote says the 1 petaflop figure is theoretical FP4 performance using the sparsity feature.
Microsoft describes a 15-inch touchscreen with a peak HDR brightness of 2,000 nits, a chassis less than 18 mm thin and under 4.5 lb, and a thermal design with up to 2.5x more thermal power capacity than its current Surface Laptops.
Microsoft is also offering up to $1,000 cash back when buyers trade in an eligible MacBook Pro, between October 7 and November 23, 2026, in the US and Canada.
Microsoft’s stated “up to” speed-ups for Windows PCs with NVIDIA RTX Spark against an Apple MacBook Pro 16-inch with M5 Pro (64 GB), from its 7 October 2026 post. Time to first token from Microsoft-commissioned testing; image and video generation from NVIDIA testing; all September 2026, preproduction PCs.US list prices (MSRP) from Microsoft’s Surface post of 7 October 2026: Surface Laptop Ultra from $2,599, available beginning October 16; Surface RTX Spark Dev Box $5,999, shipping in November.
The Wider Line-Up
Microsoft says RTX Spark Windows PCs from ASUS, Dell, HP, Lenovo and MSI are available for pre-order and begin shipping October 16. It says RTX Spark dev boxes and mini desktops will come later this year.
Against an Apple MacBook Pro 16-inch with M5 Pro and 64 GB of memory, Microsoft lists up to 2.1x faster time to first token, 4.3x faster AI image generation and 6.2x faster AI video generation. Its footnotes say the first comparison comes from Microsoft-commissioned testing and the other two from NVIDIA testing, all in September 2026 on preproduction machines.
At the top of the range, Microsoft says PCs powered by NVIDIA DGX Station will be available on Windows later this year, with up to 748 GB of coherent memory and 20 petaflops of FP4 AI compute according to its footnote.
The Software Pitch: Hybrid Intelligence
Microsoft calls its strategy hybrid intelligence: agents and models run locally when it makes sense and reach the cloud when needed. It writes that customers’ needs are outpacing what their cloud budgets can support.
It says MAI Code 1.1 Flash, a model with 137 billion total and 6.8 billion active parameters, will run on device, using 3-bit precision to cut the model size by nearly 80% while supporting a 256K context window. It also names DeepSeek V4 Flash, a 284B parameter model, and an upcoming NVIDIA Nemotron model for local use on RTX Spark.
GitHub HydraFusion, which routes each task to a model, will be able to use models running locally on Windows, with an experimental preview in the GitHub Copilot app, Copilot CLI and Visual Studio Code later in October, according to the post.
In an on-stage demo during the event, Microsoft’s Kayla Cinnamon showed a GitHub Copilot session that used a cloudmodel alongside three local sub-sessions running MAI Code 1.1 Flash Local. For one local sub-session, the on-screen panel showed 1.6M tokens sent to the model and 10.5K coming back, and she said: “All of this was free.”
Microsoft Execution Containers, which let organisations set which files and networks an agent can access, are now generally available on Windows 11. Microsoft lists Codex from OpenAI, GitHub Copilot and NVIDIA’s OpenShell among agents that support them, and Anthropic’s Claude Code among those that will.
The Installed Base
Microsoft says more than 2 trillion inferences are run locally per month across Copilot+ PCs, and that over 40% of laptops being built for business are Copilot+ PCs. It says Copilot features using local context, actions and models are expected to begin rolling out on Copilot+ PCs in the coming months.
The Structural Read
The price sets the entry point. Microsoft lists the Surface Laptop Ultra from $2,599 and the Dev Box at $5,999, and five other PC makers have RTX Spark machines on pre-order for the same October 16 ship date.
The speed claims come with conditions. Microsoft’s footnotes say the MacBook comparisons used preproduction PCs with 64 GB of memory, in September 2026, partly in NVIDIA’s own testing.
The software is meant to decide where work runs. Microsoft says HydraFusion will route tasks to models running locally on Windows, and that Execution Containers set which files and networks an agent can reach.
Microsoft, Windows Experience blog, 7 October 2026
“As AI models grow in size and capability, customers’ needs are outpacing what their cloud budgets can support.”
Three Implications
A PRICE FOR LOCAL AI Microsoft lists the Surface Laptop Ultra from $2,599 and the Surface RTX Spark Dev Box at $5,999.
MODELS SIZED FOR THE PC Microsoft says MAI Code 1.1 Flash runs on device at 3-bit precision, nearly 80% smaller.
AGENTS WITH FENCES Microsoft Execution Containers are now generally available on Windows 11, Microsoft says.
The framework’s starting point: “It is simultaneously the unit of cognition the model produces, the unit of compute the data center serves, the unit of price the lab charges, and the unit of value the enterprise extracts.”
As we read it, Microsoft’s hybrid intelligence pitch moves the second of those roles: for some tasks the token is served by a $2,599 laptop the customer owns rather than by a data center, and Microsoft’s own phrase for the benefit is “helping your tokens go further.”
What Is Not Established
We read Microsoft’s Surface and Windows posts in full. The performance comparisons are Microsoft’s and NVIDIA’s own testing on preproduction PCs, which we did not test, and the posts we read do not give prices for the partner PCs or for the top Surface Laptop Ultra configuration.
We did not contact Microsoft or NVIDIA.
Business Engineer Framework
Tokenomics: The Economics of AI
A Business Engineer framework on the token as the unit of compute, price and value across the AI stack.
Microsoft priced the Surface Laptop Ultra from $2,599, available October 16, and the Surface RTX Spark Dev Box at $5,999, shipping in November, both on NVIDIA’s RTX Spark. It is pitching them as part of hybrid intelligence, with models such as MAI Code 1.1 Flash running on device and the cloud used when needed.
94,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. We read Microsoft’s Surface and Windows posts of 7 October 2026 in full, including their footnotes. Performance figures are Microsoft’s and NVIDIA’s own testing. We did not test the devices or contact Microsoft or NVIDIA. Nothing here is a forecast, and nothing here is financial or investment advice.
Gennaro is the creator of FourWeekMBA, which reached about four million business people, comprising C-level executives, investors, analysts, product managers, and aspiring digital entrepreneurs in 2022 alone | He is also Director of Sales for a high-tech scaleup in the AI Industry | In 2012, Gennaro earned an International MBA with emphasis on Corporate Finance and Business Strategy.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
