When Satya Nadella takes the stage in San Francisco on Wednesday, he’s expected to show off the Surface Laptop Ultra, the latest step in Microsoft’s push to bring AI directly into the PCs it sells. He’ll be joined by Pavan Davuluri, executive vice president overseeing Windows and devices, and Nvidia CEO Jensen Huang, whose RTX Spark chip powers the new laptop. According to the invitation, the three are expected to discuss how “local AI will shape the next chapter of the PC.”
Wednesday won’t be the first time we see this hardware. Microsoft first introduced the Surface Laptop Ultra at Computex in June. Nadella briefly mentioned it at the start of his Build keynote, describing it as part of the company’s effort to bring “unmetered intelligence to every desk and every home.” This means that AI is available directly on device, leveraging the compute already in PCs and other edge hardware so models and agents can run locally without sending every task to the cloud.
Here’s what we know so far about the Surface Laptop Ultra:
- Price (expected well above $2,000 to start)
- Ship date (Microsoft says only “later this year”)
- Native resolution and refresh rate
- SSD capacities
- Battery capacity (Wh)
- Webcam, Wi-Fi, and Bluetooth specs
The AI Economy will be at Microsoft’s Windows and Surface event in person, live-blogging what we hear—read my coverage at the bottom of this post.
But Don’t Call it a Copilot+ PC
Microsoft’s Surface Laptop Ultra arrives during a transition at the company. It’s easy to think of the laptop as the next evolution of the Copilot+ PC, the category Microsoft invented in 2024 to define what an AI-ready Windows computer should be able to do. However, the company has moved away from the Copilot+ PC branding, and none of the Surface devices launched this year carry it. Still, Brett Ostrum, Microsoft’s corporate vice president for Surface devices, told Windows Central in an interview that the company continues to “lean into the narrative around AI on the edge and being able to have a hybrid solution out there” and that devices continue to meet the requirements of what Copilot+ devices are.
Is this launch considered a do-over for the AI PC? The promise of the Surface Laptop Ultra is a markedly different approach from when Microsoft first introduced its lineup of Copilot+ PCs—that one was immediately criticized over privacy concerns. Even before the devices shipped, the company had to address fears that its Recall feature was a security nightmare. Signal’s President Meredith Whittaker called it “a dangerous honeypot for hackers.” Two years later, Microsoft appears to have learned from that rollout. Based on what Microsoft has shown so far, the Surface Laptop Ultra looks like a steadier start, since the pitch centers on how much AI work the machine can handle on its own and leaves out the kind of headline consumer feature that gave critics an easy target in 2024.
The Surface Laptop Ultra also arrives as Microsoft reshapes Copilot, its AI assistant that will likely be the main way people use the laptop’s computing power. In the past couple of months, the company combined its two Copilot versions and created a new “super app” that puts many tools behind a paywall. Microsoft has also shuttered several consumer-centric features, suggesting Copilot is now mostly suited to prosumers. One of the more significant features included in this revamped Copilot is Autopilot—formerly Scout. It lets users create their own digital teammate, assign a role, and set a goal. The agent goes off to tackle work without being prompted. An agent that keeps working on its own for long stretches is the kind of workload the Ultra seems built for, so one thing to watch Wednesday is whether Microsoft lets Autopilot—or any OpenClaw or OpenClaw-like agent—run on the laptop’s own hardware.
As Microsoft is pushing forward with its campaign to bring “unmetered intelligence” to everyone on every device, the Surface Laptop Ultra may show users, specifically those in the enterprise, what AI can really do. Internet access is no longer a prerequisite for using the technology. And not every task should trigger anxiety over token usage. If Microsoft makes its case on Wednesday, the PC it describes will be less a tool people prompt and more a place where agents carry out assigned work on their own, using computing power that’s already sitting on the desk.
Support for On-Device AI Models
Microsoft is pitching the Surface Laptop Ultra as something more than your ordinary laptop. If it was, the company could have simply announced it with a press release or blog post. What’s the technology that makes this perhaps the epitome of an AI PC? People tend to use cloud-based AI for their tasks, but if the Surface Laptop Ultra is built to create Frontier Professionals, it needs to make AI available in all situations, and that includes locally on the device. Will Microsoft address how its first-party models such as those created by Microsoft AI—MAI-Transcribe, MAI-Thinking, MAI-Voice, MAI-Code-Flash—and others such as Aion Instruct and Aion Plan will make the Surface Laptop Ultra stand out?
How Much Will the Surface Laptop Ultra Cost?
While the compute power is enough to make any AI-focused builder salivate, one key question remains unanswered: how much will the Surface Laptop Ultra cost? Expect Microsoft to provide that information on Wednesday, but don’t be surprised that it’s not going to be cheap. Everything has become more expensive, technology included. RAM prices have spiked due to a severe shortage. SSD prices have also soared. Computer hardware companies have raised their prices, including Apple, Dell, Lenovo, Asus, Framework, and Microsoft.
Microsoft calls the Surface Laptop Ultra its most powerful Surface laptop yet, and machines in that class rarely come cheap. That may suggest a price well above the roughly $1,200 that the standard Surface Laptop sells for today. Apple’s MacBook Pro, the laptop the Ultra has most often been measured against, starts at $1,999, which offers a reasonable benchmark for where Microsoft might aim.
That said, Microsoft has also released more budget-friendly options, just in time for the holiday shopping season. In late September, the company introduced its next-generation Surface Pro 12-inch and Surface Laptop 13-inch. What sets them apart is that they’re now powered by Qualcomm’s Snapdragon X2 Plus processor. They start at $1,149.99 and $1,199, respectively. But while both the Surface Pro and Surface Laptop can run AI locally, they’re built more for everyday use and lighter local tasks. The Surface Laptop Ultra is designed for developers and power users who want to run large models on the device itself.
What Will Nvidia Reveal About RTX Spark?
It seems that Huang will do more than make a guest appearance at Microsoft’s Windows and Surface event. Reportedly, Nvidia may use the opportunity to share more about the RTX Spark N1X chip that is slated to arrive in RTX Spark laptops, one of which is Microsoft’s Surface Laptop Ultra. RTX Spark is described as a PC-first platform built to run games, produce content, and execute AI tasks. At IFC in September, Nvidia announced that the first computers to support RTX Spark would become available in October.
Could we also hear more about the RTX Spark Dev Box that Microsoft unveiled at Build in June? Nadella said at the time that it would be available in the fall, so what better time would there be to share more information about it than at an event discussing the Surface Laptop Ultra that’s powered by Nvidia’s RTX Spark chip?
Disclosure: I’m attending Microsoft's Windows and Surface Event as a guest of the company, with travel and accommodation expenses covered. What I write reflects my own reporting and analysis, and no one reviewed or approved this post before publication.
Microsoft Launches the Surface Laptop Ultra and RTX Spark
EndedMicrosoft CEO Satya Nadella, Executive Vice President for Windows and Devices Pavan Davuluri, and Nvidia CEO Jensen Huang will speak at a special Windows and Surface event on Oct. 7 in San Francisco about the latest updates to Windows, RTX Spark, Surface, and the broader PC ecosystem.
Coverage has ended
Discover more from Ken Yeung
Subscribe to get the latest posts sent to your email.
