Microsoft Unveils New Surface Laptop Ultra Built for Local AI
Microsoft has introduced a new generation of high-performance Windows computers designed to run powerful artificial intelligence workloads directly on personal devices.
The company has unveiled the Surface Laptop Ultra, a premium 15-inch laptop powered by NVIDIA's new RTX Spark Superchip. Microsoft is positioning the device as its most powerful Surface laptop yet, with a major focus on local AI, advanced graphics and AI-agent workloads.
The new laptop starts at $2,599, with pre-orders now open and availability beginning October 16, 2026.
NVIDIA RTX Spark Powers the New Surface
The Surface Laptop Ultra is built around NVIDIA's RTX Spark platform, which combines a Blackwell-based RTX GPU with an NVIDIA Grace CPU.
The system can be configured with up to 128GB of unified memory and up to 6,144 GPU cores, giving the machine substantially more computing capability than a conventional consumer laptop.
Microsoft says the system can deliver up to 1 petaflop of AI performance and can run AI models exceeding 120 billion parameters locally.
This means developers and other users can perform demanding AI workloads directly on the computer instead of sending every task to a remote cloud server.
Microsoft Wants AI to Run Both Locally and in the Cloud
Microsoft is calling its new approach hybrid intelligence.
Under this model, Windows can determine whether an AI task should be processed locally on the PC or sent to Microsoft's cloud infrastructure.
Local processing can provide benefits such as lower dependence on cloud services, greater control over sensitive data and potentially lower costs for workloads that can run efficiently on the device.
Cloud computing can still be used when a task requires significantly more processing power than the local machine can provide.
Microsoft believes this combination could become an important part of the next generation of AI-powered personal computers.
AI Models Can Run Directly on the Laptop
One of the biggest changes is Microsoft's push to make powerful AI models practical on personal computers.
The company says models including its MAI Code 1.1 Flash, an upcoming NVIDIA Nemotron model and DeepSeek V4 Flash can be optimized to run locally on RTX Spark systems.
Microsoft says MAI Code 1.1 Flash has 137 billion total parameters and 6.8 billion active parameters. Using 3-bit precision, the model can be reduced substantially in size while maintaining its coding capabilities.
The company says the model can operate locally with a 256K context window.
New Security System for AI Agents
Microsoft is also addressing one of the biggest challenges associated with autonomous AI agents: security.
AI agents can potentially read files, use software tools, write code and interact with networks without requiring a human to approve every individual action.
That creates a new category of security risks because an agent with excessive permissions could potentially access information it should not be able to reach.
Microsoft has therefore made its Microsoft Execution Containers generally available on Windows 11.
Microsoft Execution Containers Control What AI Agents Can Access
Microsoft Execution Containers are designed to provide a controlled environment for AI agents.
Organizations can define which files and networks an agent is allowed to access, while the system enforces those policies during operation.
Microsoft describes the system as part of a broader security model based on three principles: containment, identity and manageability.
Containment limits what an agent can access. Identity helps organizations distinguish actions performed by an AI agent from actions performed by a human user. Manageability gives IT teams tools to monitor and govern agent activity.
OpenAI and Anthropic Among the Companies Supporting the Technology
Microsoft says several AI tools and platforms are already supporting Microsoft Execution Containers.
These include OpenAI Codex, GitHub Copilot, OpenClaw, Replit, LM Studio, NVIDIA OpenShell and Unsloth AI.
Microsoft also says support is coming for additional AI systems, including Anthropic Claude Code, Box, Egnyte, Perplexity, Raycast and other agent platforms.
This could make Windows an increasingly important operating environment for developers building autonomous AI applications.
Surface Laptop Ultra Is Designed for AI Developers
Microsoft is targeting the new laptop at developers, creators, AI engineers and other professionals who require substantial local computing power.
The company says developers can use local models with GitHub Copilot to perform coding tasks directly on the device.
This can allow developers to prototype AI agents, review code, test larger models and perform other development tasks without relying entirely on cloud infrastructure.
The laptop also supports NVIDIA's CUDA ecosystem, allowing developers to use a broad range of existing AI software and tools.
Surface RTX Spark Dev Box Targets AI Engineers
Alongside the laptop, Microsoft has announced the Surface RTX Spark Dev Box.
This is a compact desktop system designed specifically for developers and AI engineers who need powerful local computing at their desks.
The Dev Box also features up to 128GB of unified memory and up to 1 petaflop of AI performance.
Microsoft says it can run AI models exceeding 120 billion parameters locally and is designed for inference, evaluation and AI experimentation.
The Surface RTX Spark Dev Box is priced at $5,999 and is scheduled to begin shipping in the United States in November.
Microsoft Is Building a Wider RTX Spark PC Ecosystem
The Surface Laptop Ultra is not the only computer using NVIDIA's RTX Spark platform.
Microsoft says manufacturers including ASUS, Dell, HP, Lenovo and MSI are also preparing Windows PCs based on RTX Spark.
This creates a broader hardware ecosystem for local AI rather than limiting the technology to Microsoft's own Surface devices.
The strategy could make powerful local AI available across a wider range of Windows computers.
Microsoft Is Also Bringing AI Supercomputers to Windows
At the higher end of the market, Microsoft says DGX Station for Windows will bring data-center-class AI computing to desktop environments later this year.
The system will use NVIDIA's GB300 Grace Blackwell Ultra Desktop Superchip and is designed to run extremely large AI models locally.
Microsoft says the platform can support frontier-class models with more than one trillion parameters.
This shows the company's broader strategy: from AI laptops for developers to workstation-class systems capable of handling workloads traditionally associated with data centers.
A New Battle for the AI PC Market
Microsoft's latest hardware strategy comes as technology companies increasingly compete to bring artificial intelligence directly onto personal devices.
Apple has been expanding its own AI capabilities across its hardware ecosystem, while Microsoft is using Windows and NVIDIA hardware to push a hybrid local-and-cloud model.
The competition is no longer only about faster processors or better displays. AI capability, memory capacity, privacy and the ability to run models locally are becoming increasingly important factors in high-end computing.
What the Surface Laptop Ultra Means for the Future
The Surface Laptop Ultra represents Microsoft's attempt to redefine what a professional Windows laptop can do.
With up to 128GB of unified memory, NVIDIA Blackwell graphics and local support for large AI models, the device is designed to perform tasks that previously required powerful desktop workstations or cloud infrastructure.
The introduction of Microsoft Execution Containers also shows that the company is preparing for a future where AI agents become regular users of computers rather than simply software applications controlled by humans.
Local AI Could Become the Next Major PC Shift
The move toward local AI could change the role of the personal computer.
Instead of simply running applications, future PCs may continuously run AI models that understand local files, assist with coding, automate repetitive tasks and interact with software on behalf of users.
However, that increased capability also requires stronger security controls to prevent AI agents from accessing information or systems without permission.
Microsoft's combination of RTX Spark hardware, local AI models and Execution Containers is designed to address both sides of that transition.
Microsoft's New AI PC Era Begins
The Surface Laptop Ultra and Surface RTX Spark Dev Box mark a significant expansion of Microsoft's local AI strategy.
The company is combining Windows, NVIDIA hardware, local AI models and new security technology to create computers capable of running increasingly sophisticated artificial intelligence directly on the device.
With the Surface Laptop Ultra starting at $2,599 and shipping from October 16, Microsoft's new AI-focused hardware will soon begin reaching customers.
As AI agents become more capable and models become more efficient, the personal computer could increasingly become an AI computing platform rather than simply a device for accessing cloud services.
Journalist: Vijay Singh