Who makes CPUs? Well, you’ve got AMD, Intel, Apple, Qualcomm, and Mediatek as the major players in consumer devices, with companies like Samsung, Broadcom, Google, Huawei, UNISOC, and Rockchip each making up smaller slices of the market (even if their portfolios are often much larger). NVIDIA’s there too, though; while everyone knows about NVIDIA’s GeForce gaming GPUs and dedicated AI chips, NVIDIA sells quite a few CPUs, too. Primarily, at this time, that’s through Nintendo’s handhelds—at least in terms of volume—but the company is clearly eyeing the laptop and desktop market as well with its upcoming N1/N1X processors.

If you’re desperate for an all-NVIDIA gaming machine, you likely won’t have long to wait. Digitimes reports that Windows devices running N1[X] processors are scheduled to launch this quarter, meaning no later than March. That’s a little surprising, if true, simply because we haven’t heard a peep about N1[X] out of anyone aside from an offhanded mention from Jensen Huang that N1X is based on the same silicon as the GB10 Superchip in the green team’s DGX Spark desktop AI workstation.

That means a 10+10 Arm CPU configuration and 48 Blackwell shader modules, resulting in 6144 CUDA cores—the very same setup as a GeForce RTX 5070. Of course, the integrated graphics on GB10 are nowhere near as fast in games as a GeForce RTX 5070, but that’s because they’re held back by a shared 256-bit LPDDR5X memory bus. This configuration can provide potent gaming performance (just look at our reviews of AMD’s Ryzen AI Max, which have the same setup), but it’s not comparable to GDDR7 discrete GPUs.

HP’s Zbook Ultra G1a is likely rather similar to the forthcoming N1X machines.

What would a laptop with a GB10 inside be like? Well, probably a lot like the HP ZBook Ultra G1a that we reviewed, which is a laptop with a Ryzen AI Max processor inside. It’s difficult to overstate how similar these two parts are: gobs of CPU cores, a huge integrated GPU, and a double-wide LPDDR5X memory bus. It’s not clear what separates the N1 from the N1X, but it’s most likely to be a matter of either cut-down resources or simply capped clock rates, not something as extreme as a major silicon switch or discrete GPU configuration.

Even still, in other ways, these systems couldn’t be further apart. The Ryzen AI Max has regular (desktop-class) Zen 5 CPU cores, so it’s compatible with all the software you would usually run on a Windows system. The N1[X] uses Arm CPU cores; it’ll be running the same version of Windows (with the same translation layer) as Qualcomm’s Snapdragon X Elite processors. We ran into no end of compatibility problems in our reviews of the first Copilot+ PCs, but it’s been more than a year since then, and things have improved considerably in that regard.

A die shot of NVIDIA’s GB10 Superchip, showing separate CPU and GPU chiplets.

Does this mean you should get excited about systems sporting the NVIDIA N1[X] SoC? Well, maybe. It’s possible that the chip could offer an impressive balance of performance and battery life. We found that AMD’s Strix Halo, the Ryzen AI Max parts discussed above, had better battery life than discrete-GPU gaming laptops, yet better performance than standard integrated graphics, leaving it in an interesting middle ground. We expect NVIDIA’s N1X will end up in a similar place, with perhaps superior power efficiency due to being fabricated on a newer 3nm process versus the 4nm process of Strix Halo.

With that in mind, ultimately it’s all going to come down to pricing. While we liked the ‘middle ground’ aspect of Ryzen AI Max, it’s honestly impossible to recommend to most buyers due to the extremely high price such systems ask. It’s quite possible that NVIDIA’s gaming SoC could fall into the same product class, leaving it an interesting curiosity rather than a serious competitor. We’ll seemingly know sooner than later.

Zak Killian

A 30-year PC building veteran, Zak is a modern-day Renaissance man who may not be an expert on anything, but knows just a little about nearly everything.