A few days ago, we shared with you our disappointment at not being able to see Nvidia’s ARM N1X chips in action when we returned from Computex. These chips are expected to arrive in desktop and notebook PCs running Windows on Arm (like Microsoft Surface PCs and their Snapdragon Soc). Promising and sometimes paradoxical performance figures were leaked, as was an exclusive launch with Alienware, presumably for a gaming notebook. But we suspected that these delays and rumors were not good news. And the announcement of a postponement to 2026 has confirmed these fears. The main reason for this postponement has not been officially communicated, but several media outlets (including us) have relayed the idea of a major design flaw forcing the greens to revise their copy in depth. But in the last few days, another version has emerged, and it makes for a completely different reading of the postponement. The big “culprit” in the delay of the N1X chips would be, in part, Microsoft.
Windows on Arm: has Microsoft oversold the baby?
Several sources quoted by Digitimes give a different reading of the delays to the chip resulting from the collaboration between Nvidia and Mediatek. The chips in question are said to be based on the GB10, the heart of the Nvidia DGX Spark mini-supercomputer. While regular modifications to the chip on the Nvidia side have been confirmed and are contributing to the delay, problems in keeping to the roadmap promised by Microsoft for its Windows on ARM are another real problem. Finally, after the hype surrounding the ARM/Windows combo, it seems that the bellows are suddenly falling again. Demand on the global notebook market shows that machines based on the Qualcomm Snapdragon / Windows on Arm pairing are less in demand than their first appearances at launch might have suggested.
We weren’t convinced by the ARM / Windows combo
Let’s be clear: our own tests of ARM / Windows machines didn’t convince us. Even in classic use, we didn’t find what we were looking for. So what about using this platform for games? Obviously, we’d like to see things that shake up our market. But the arrival of ARM chips under Windows was justified by the desire for chips that would increase autonomy. Since then, Intel’s Lunar Lake has shown that we don’t need ARM. What are the reasons today for persevering with a solution that imposes more constraints than benefits?