Microsoft and NVIDIA are pushing Windows PCs into a new phase of local artificial intelligence with the launch of the Surface Laptop Ultra, powered by NVIDIA’s new RTX Spark platform. The companies unveiled the new hardware and a series of Windows AI improvements at an event in San Francisco, positioning the devices as computers designed to run demanding AI workloads directly on the machine.
The new Surface Laptop Ultra is aimed at developers, creators and AI builders who need considerably more local computing power than conventional laptops can provide. It combines NVIDIA’s Blackwell-based RTX Spark technology with up to 128GB of unified memory and up to one petaflop of AI performance.
Microsoft says the laptop can run AI models with more than 120 billion parameters locally, reducing the need to send every demanding AI workload to the cloud.
Surface Laptop Ultra Price and Availability
The Surface Laptop Ultra starts at $2,599 in the United States, with higher configurations increasing the price depending on the processor, memory and storage options.
Microsoft has opened pre-orders, with the first devices scheduled to become available on October 16, 2026. The RTX Spark-based Surface lineup is being positioned as Microsoft’s most powerful Surface hardware to date.
The laptop is also being joined by the Surface RTX Spark Dev Box, a compact developer-focused system designed for users who need powerful local AI computing at a desk. The Dev Box starts at $5,999 in the US, with shipments expected to begin in November.
What Is NVIDIA RTX Spark?
At the heart of the new Surface Laptop Ultra is NVIDIA’s RTX Spark platform.
RTX Spark combines a powerful NVIDIA Blackwell RTX GPU with a high-performance Grace CPU in a unified architecture. The configuration can include up to 6,144 GPU cores and a 20-core CPU, connected through a high-speed memory architecture.
The platform is designed to deliver up to one petaflop of FP4 AI performance and support up to 128GB of unified memory. Unlike traditional PCs where CPU and GPU memory are generally separated, the unified-memory approach allows workloads to dynamically use the available memory across the processor and graphics hardware.
That is particularly important for AI applications because larger models can require huge amounts of memory even before they begin processing data.
Designed to Run Large AI Models Locally
The biggest selling point of the Surface Laptop Ultra is its ability to run large AI models locally.
Microsoft says the system can handle models exceeding 120 billion parameters directly on the laptop. This could allow developers and advanced users to experiment with models locally instead of relying entirely on cloud-based AI services.
Local processing can have several advantages. Sensitive data can remain on the user’s machine, while developers can also avoid sending every request to a remote server.
There can also be latency and cost benefits. A local model does not require an internet connection for every inference request, while developers testing AI applications can experiment without constantly paying for cloud compute.
Of course, cloud AI will continue to be useful for workloads that exceed the capabilities of a single PC. Microsoft’s broader strategy is therefore based on hybrid intelligence, allowing workloads to move between local hardware and the cloud depending on what is required.
Up to 128GB of Unified Memory
Memory is one of the most unusual aspects of the Surface Laptop Ultra.
The system can be configured with up to 128GB of unified memory, giving AI applications considerably more room to work with large models and datasets.
Microsoft says the memory architecture can dynamically allocate resources between CPU and GPU workloads. This can be particularly useful when an application needs both traditional processing and GPU acceleration at the same time.
For creators, the additional memory could also help with demanding workflows such as high-resolution video editing, 3D rendering and AI-assisted content generation.
A 15-Inch Surface Designed for Heavy Workloads
The Surface Laptop Ultra features a 15-inch touchscreen and a new thermal architecture designed to maintain performance during sustained workloads.
That is important because AI workloads can keep a processor under heavy load for extended periods. Rather than optimizing only for short bursts of performance, Microsoft says the new thermal design is intended to support demanding workloads over longer sessions.
The device therefore represents a significant departure from the lightweight productivity-focused Surface laptops Microsoft has traditionally offered.
This is a Surface machine designed specifically for people who need more computing power.
RTX Spark Is Not Just for AI
Although AI is the main focus, NVIDIA is also positioning RTX Spark as a platform for creators and gamers.
The RTX Spark architecture includes NVIDIA’s CUDA ecosystem, fifth-generation Tensor Cores, RTX ray tracing and DLSS technologies.
NVIDIA says RTX Spark systems can support demanding gaming workloads, including AAA games at 1440p with frame rates exceeding 100fps in supported configurations and technologies.
Creators can also benefit from hardware-accelerated video processing, AI-powered editing and advanced rendering.
That makes RTX Spark different from a traditional AI accelerator. It is designed to combine AI, graphics, development and gaming capabilities into one PC platform.
Microsoft Brings AI Agents to Windows
The hardware launch is only part of Microsoft’s announcement.
The company also introduced new Windows capabilities designed around AI agents.
Microsoft is moving its Microsoft Execution Containers technology out of beta and into general availability. The system is designed to provide a controlled environment where AI agents can operate while limiting their access to local files, networks and tools.
This is important because AI agents are increasingly capable of taking actions rather than simply generating text.
An agent that can modify files, run commands or interact with applications needs stronger security boundaries than a conventional chatbot.
Microsoft is therefore positioning Windows as an operating system designed not only for people but also for AI agents working alongside them.
GitHub Copilot Gets More Local AI Capabilities
The new RTX Spark hardware is also being integrated with Microsoft’s developer ecosystem.
Microsoft demonstrated AI coding workloads involving GitHub Copilot, including workflows where AI agents can analyze builds and suggest fixes.
The additional local computing power means developers can potentially run more demanding AI-assisted coding workloads directly on their PCs.
Microsoft is also bringing additional local models and tools to Windows, strengthening its effort to make the PC a practical platform for AI development rather than simply a client for cloud-based AI services.
RTX Spark PCs From Other Manufacturers
Microsoft isn’t keeping RTX Spark exclusive to Surface.
NVIDIA says systems powered by the platform are coming from manufacturers including ASUS, Dell, HP, Lenovo, MSI and Microsoft, with additional models from Acer and Gigabyte also part of the broader ecosystem.
This could make RTX Spark more important than a single Surface product.
If multiple manufacturers adopt the architecture, consumers and developers will eventually have a broader range of laptops and desktops designed around local AI.
The systems are expected to range from relatively portable laptops to compact desktop machines designed for always-on AI workloads.
Surface RTX Spark Dev Box for Developers
Alongside the laptop, Microsoft introduced the Surface RTX Spark Dev Box.
The machine is designed specifically for developers and AI engineers who need a dedicated local AI system rather than a traditional laptop.
Microsoft says it can provide powerful local compute for experimenting with models, evaluating AI applications and building agent-based workflows.
The Dev Box is priced significantly higher than the Surface Laptop Ultra, reflecting its more specialized purpose.
It is expected to begin shipping in the US in November.
NVIDIA DGX Station Comes to Windows
For organizations that need even more AI performance, NVIDIA and Microsoft also announced DGX Station for Windows.
The system uses NVIDIA’s GB300 Grace Blackwell Ultra Desktop Superchip and is designed to bring data-center-class AI computing to a workstation environment.
NVIDIA says the system offers 748GB of coherent memory and up to 20 petaflops of FP4 AI performance, allowing it to handle models at an entirely different scale from conventional PCs.
The idea is to allow enterprise AI teams to run and develop very large models locally while remaining inside the Windows ecosystem.
Why the Surface Laptop Ultra Matters
The Surface Laptop Ultra is important because it represents a shift in what Microsoft believes a premium PC should do.
Traditional laptops have largely been built around applications such as web browsers, office software, creative programs and games. AI has increasingly become another workload, but much of the most demanding processing still happens in the cloud.
Microsoft and NVIDIA are attempting to change that balance.
With RTX Spark, large amounts of unified memory and dedicated AI hardware, the new Surface Laptop Ultra is designed to make local AI practical for workloads that previously required expensive cloud infrastructure.
The approach could also give users more control over their data.
Final Verdict
The Microsoft Surface Laptop Ultra is more than a new premium Surface laptop. It is Microsoft’s first major showcase for a new generation of Windows PCs built around NVIDIA RTX Spark and the idea of hybrid intelligence.
With up to 128GB of unified memory, one petaflop of AI performance and support for models exceeding 120 billion parameters, the machine is clearly aimed at demanding users rather than mainstream laptop buyers.
The broader RTX Spark ecosystem could ultimately be even more significant. With systems from multiple PC manufacturers, developer-focused machines and NVIDIA’s DGX Station for Windows, the companies are building an entire hardware category around local AI.
For now, the Surface Laptop Ultra starts at $2,599, with pre-orders already open and availability beginning October 16. The launch marks a significant step toward Microsoft’s vision of Windows PCs that can work alongside AI agents while keeping more powerful AI workloads directly on the device.
What is the Microsoft Surface Laptop Ultra?
The Surface Laptop Ultra is Microsoft’s new high-performance AI laptop powered by NVIDIA’s RTX Spark platform, designed for demanding local AI, development, creative, and gaming workloads.
What is the price of the Surface Laptop Ultra?
The Microsoft Surface Laptop Ultra starts at around $2,599 in the US, with higher-end configurations costing more.
Can the Surface Laptop Ultra run AI models locally?
Yes. Microsoft says the RTX Spark-powered laptop can run AI models with more than 120 billion parameters locally, reducing the need to rely entirely on cloud-based processing.
When will the Surface Laptop Ultra be available?
Microsoft has announced that the Surface Laptop Ultra will begin shipping on October 16, 2026, following preorders.
What is NVIDIA RTX Spark?
NVIDIA RTX Spark is the platform that powers Surface Laptop Ultra and serves as the foundation for its graphics, AI, and performance capabilities.
It combines:
Blackwell RTX graphics for advanced AI capabilities, visual computing, and graphics performance
Up to a 20-core CPU for demanding workflows and multitasking
Up to 128 GB of unified memory for large files, datasets, and AI workloads
Does Surface Laptop Ultra have a touchscreen?
Yes. Surface Laptop Ultra includes a 15-inch PixelSense™ Ultra touchscreen that supports direct interaction with your work. Zoom, mark up, scroll, sign, and navigate more naturally, whether you’re reviewing footage, refining creative work, exploring data, or collaborating on a project.
