Surface Laptop Ultra
Surface Laptop Ultra: The Most Powerful Surface Ever. But That’s Not the Story.
Microsoft’s newest Surface isn’t about faster processors or bigger specifications. It’s about the next phase of AI: moving intelligence from the cloud onto the device, while combining it with the scale of the cloud when needed.
Microsoft calls Surface Laptop Ultra the most powerful Surface Laptop ever. Look at the specifications and it’s easy to see why. But if you stop at the spec sheet, you’ll miss what this device tells us about where business computing is heading.
The real story is where AI runs.
Three eras of Microsoft AI
To see why Surface Laptop Ultra matters, it helps to look at how Microsoft’s AI story has developed over the last few years.
Era one: the cloud. AI lived in Azure. If you wanted serious AI capability, you rented it from a data centre, paid for it by consumption, and sent your data over the network to use it.
Era two: Copilot. AI moved into the tools people use every day. Microsoft 365 Copilot brought generative AI into Outlook, Teams, Word and Excel. AI stopped being a ‘specialist’ service and became part of everyday work, but it still ran almost entirely in the cloud.
Era three: hybrid AI. This is where we are now. Copilot+ PCs added a Neural Processing Unit (NPU) to the everyday laptop, so features like Live Captions, Windows Studio Effects and semantic search run on the device itself. No cloud round trip, and in many cases no Copilot licence is needed.
Surface Laptop Ultra takes that third era a big step further.
From AI features to AI workloads
A standard Copilot+ PC runs AI features locally. Surface Laptop Ultra is designed to run AI workloads locally.
It’s built around NVIDIA’s RTX Spark superchip, which combines an NVIDIA Blackwell RTX GPU, an NVIDIA Grace CPU and up to 128 GB of unified memory shared between the two. Microsoft says it can run AI models with more than 120 billion parameters locally, with up to 1 petaflop of AI performance.
Put simply, work that used to need a server; a specialist workstation or a cloud subscription can now run on a 15-inch laptop that fits in a backpack.
Microsoft also tackled the problem that usually holds back powerful thin laptops: heat. A new thermal design delivers up to 2.5 times the thermal capacity of current Surface Laptops, in a chassis under 18 mm thin, so the device can sustain heavy work rather than just reach peak performance briefly.
Why where AI runs matters to your business
For most organisations, local AI isn’t about speed for its own sake. It comes down to three practical questions.
Control. When a model runs on the device, the data, the prompt, and the answer can stay on the machine. For sensitive client records, intellectual property or confidential financial data, that changes the risk conversation.
Cost. Cloud AI is metered. Every prompt, every test and every agent step costs something. Microsoft makes this point directly: teams can experiment locally, cut their reliance on the cloud, and save cloud resources for the workloads that really need them.
Connectivity. In Western Australia, a lot of real work happens far from fast, reliable internet: mine sites, regional offices, field crews and remote projects. AI that keeps working offline isn’t a nice-to-have here. It’s a requirement.
Hybrid AI doesn’t replace the cloud. It puts each workload where it makes the most sense: on the device for private, repeatable and offline work, and in the cloud for frontier-scale tasks.
Governed, not just powerful
The IT question always follows the excitement: how do we keep control of this?
Microsoft is building governance into the local AI story as well. Microsoft Execution Containers provide containment for AI agents. Upcoming Windows capabilities will use Microsoft Entra to tell agent activity apart from user activity and will extend Microsoft Agent 365 controls to agents running locally on the device.
That matters. Powerful local AI without governance would just be shadow IT with a faster GPU.
So, who is it actually for?
To be honest: probably not most of your staff.
If someone’s day is email, Teams, documents and spreadsheets, a standard Surface Pro or Surface Laptop Copilot+ PC is the right tool. It runs on-device AI features well and is far easier on the budget.
Surface Laptop Ultra is for the people Microsoft designed it around: developers, creators, data scientists and AI builders. In practice, that could include:
- Developers running local models with GitHub Copilot, prototyping agents and testing larger models without a cloud bill for every attempt
- Creative and media teams needing colour accuracy, GPU acceleration and a display suitable for professional HDR work, with a quoted peak brightness of 2,000 nits
- Engineering and technical teams working with large datasets, models and visualisations, often away from the office
- Innovation teams exploring what private, on-device AI could mean for the organisation before committing to cloud spend
The smart approach isn’t one device for everyone. It’s matching the device to the role.
The bigger takeaway
Surface Laptop Ultra is impressive hardware, but it’s also a signal. The endpoint matters again.
For the past decade, the laptop has been increasingly seen as a window into the cloud. In the hybrid AI era, the device you choose decides what AI your people can run, where their data goes, and how much each workload costs. That makes device strategy a core part of AI strategy.
Even if Surface Laptop Ultra isn’t right for most of your team, the question it raises is right for every organisation: where should your AI run?
Talk to Stott Hoare Whether you’re planning your next device refresh, building an AI roadmap, or wondering which roles justify AI-ready hardware, our team can help you match the right device to the right work.
