At its Windows event on October 7, Microsoft launched the Surface Laptop Ultra with Nvidia’s RTX Spark chip. It’s decent hardware, but the company also used the launch to make some bold claims about the MacBook Pro. According to Microsoft, its new RTX Spark-powered Windows PCs are up to 6.2x faster than the M5 Pro-powered MacBook Pro, but it’s not what you think.
Microsoft wants you to believe that the Surface Laptop Ultra and other Windows PCs with Nvidia RTX Spark are better choices than Apple’s hardware for local AI.
“The new Surface Laptop Ultra is built for this new era of creative work,” Pavan Davuluri, head of Windows and Devices at Microsoft, said at the event.
“Across the models we tested, Surface Laptop Ultra delivers faster time-to-first-token performance for language models and shorter image generation time than M5 Pro-powered laptops.”
According to Microsoft’s benchmarks from the event, Surface Laptop Ultra with RTX Spark offers:
- up to 2.1x faster time to first token
- up to 4.3x faster AI image generation
- up to 6.2x faster AI video generation
The tests compared selected preproduction Surface Laptop Ultra RTX Spark PCs with 64GB of memory against a 16-inch MacBook Pro with an M5 Pro and 64GB. Surface Laptop Ultra is specifically named in the time-to-first-token test, while the image and video tests refer to a selection of RTX Spark PCs.
However, it’s worth noting that Microsoft is only hyping AI numbers, so it doesn’t plan on telling us how the two laptops compare in everyday tasks, and that includes non-AI video editing, overall performance, and battery life.
Microsoft is maintaining radio silence over any other comparisons with the M5 Pro.
“This is where having both the large memory and the petaflop of AI compute on a single device is a distinct advantage,” Davuluri added. “For builders using local AI, that means less waiting and more creating on Windows.”
I watched the rest of the event, which started to bore me to death at one point, but Microsoft didn’t follow up with a head-to-head CPU benchmark for RTX Spark and Apple’s M5 Pro.
Microsoft wants consumers and “builders” to buy Surface Laptop Ultra for local AI
Surface Laptop Ultra is an interesting device, and it’s largely built around the idea of running AI locally.
Microsoft may have realized that it needs to do more to compete with Apple in the local AI race. As you might remember, hundreds of thousands of people rushed to buy a Mac just to run OpenClaw when it was trending. We haven’t seen that kind of “craze” to buy a Windows PC to run AI agents locally, and Microsoft wants that to change because it feels threatened.
This is where the Surface’s original vision comes into the picture.
For those unaware, Microsoft builds different Surface products for different reasons, and it expects its OEM partners to follow that direction. Microsoft doesn’t sell Surface products just to make revenue, and you’d have noticed this already if you’ve been following the company’s hardware strategy.
Surface is part of the company’s broader strategy to “push” and “inspire” OEMs to innovate. When Microsoft launched Surface Book in 2015, then-hardware chief Panos Panay called it the ultimate laptop. And during the launch event, Microsoft targeted the MacBook Pro, rather than the likes of Lenovo or HP.
We’ve seen the same approach more recently with Copilot+ PCs powered by Snapdragon processors. At that point, the MacBook comparison was around processor performance, not really AI. Microsoft likes to go after the MacBook when it has an idea to sell, and this time, that idea is local AI.
Microsoft calls Surface Laptop Ultra its most powerful Surface hardware and says “it took end-to-end engineering, design, and full-stack thinking to unlock what local AI can do.”
With up to 128GB of unified memory and up to a petaflop of AI compute, Microsoft is hoping that “builders” will pick up Surface Laptop Ultra to run models that don’t fit on a traditional laptop.
The company also claims up to 2.5 times the thermal capacity of its existing laptops and said at the event that the device can maintain “over 99% of its power” when running on battery.
