What Nvidia confirmed
The first RTX Spark Windows PCs are scheduled to arrive in October, according to Nvidia’s September 3 announcement and a report by The Verge published the same day. Nvidia’s official specification page lists two N1X configurations: one combines a 20-core Grace CPU with a 6,144-core Blackwell RTX GPU, while the other uses an 18-core Grace CPU and a 5,120-core GPU. Both target Windows 11 systems and operate in a configurable 45–80 watt range.
The platform uses unified memory rather than separate system and graphics pools. Nvidia lists up to 128 GB for the larger configuration and up to 64 GB for the smaller one. That design is especially relevant to local AI workloads, where model size can be constrained by available accelerator memory.
Why the split matters
The two versions give manufacturers room to balance thin-laptop thermals, memory capacity and graphics resources. The higher configuration can also appear in compact desktops, while the reduced version is aimed at laptops. Nvidia has named systems from Asus, Dell, HP, Lenovo, Microsoft and MSI on its product page, but final specifications will depend on each manufacturer.
This is not yet independent performance evidence. Nvidia’s claims about AI, creative and gaming capability come from the platform vendor; buyers should wait for shipping systems, battery tests and application-compatibility reviews before comparing RTX Spark with x86 or other Arm laptops.
What to watch in October
The practical questions are software compatibility on Windows on Arm, sustained performance inside thin designs, and how much unified memory vendors include in retail configurations. Pricing was not part of Nvidia’s September 3 availability announcement. The launch will therefore be less about the headline core counts than about the complete systems that manufacturers actually ship.