NVIDIA has put a date on its most consumer-facing hardware move yet. A teaser posted by Microsoft's official Windows account on X - and confirmed by NVIDIA's own RTX Spark account - pins the launch of the RTX Spark platform to October 7, where it will be shown off at a Windows and Surface event in San Francisco. It is the first time NVIDIA has designed, branded and launched an entire consumer PC platform rather than selling chips into someone else's design.
At the heart of the platform is the N1X system-on-chip, which fuses a Grace CPU with an RTX GPU built on NVIDIA's Blackwell architecture into a single package. Both sides of the chip share one pool of LPDDR5X memory rather than splitting system RAM from video memory, and the silicon is paired with software NVIDIA has already shipped for it: RTX Kit 2026.3 and version 2.0 of the RTX Mega Geometry SDK, which bring neural rendering and path-tracing tools to developers building for the platform.
Two configurations have been disclosed. The flagship combines a 20-core Grace CPU with a Blackwell GPU carrying 6,144 CUDA cores, addresses 24GB to 128GB of unified memory, and is intended for both laptops and mini PCs. The second tier pairs an 18-core Grace CPU with 5,120 CUDA cores and caps memory at 24GB to 32GB; those machines are laptop-only and will be classified as Copilot+ AI PCs running Windows 11 on Arm. NVIDIA says the platform supports DLSS, Reflex and Game Ready drivers.
The launch partner list is unusually broad for a first generation: Acer, ASUS, Dell, Gigabyte, HP, Lenovo, Microsoft and MSI are all confirmed. The first wave of machines, however, leans heavily toward professionals. Six laptops have been announced so far - ASUS ProArt P16, Dell XPS 16, HP OmniBook Ultra 16, Lenovo Yoga Pro 9n, Microsoft Surface Laptop Ultra and MSI Prestige N16 Flip AI+ - and not one of them is a gaming laptop, despite RTX branding that is synonymous with gaming hardware.
That positioning matches how NVIDIA describes the platform: personal AI compute, content creation and professional graphics, with the large unified memory as the selling point for running models locally. ASUS has been warming up its ProArt 16 ahead of an October 8 debut, and MSI's Prestige N16 Flip AI+ opens pre-orders the same day.
Two caveats are worth holding onto before the keynote. First, most of the six machines have no published prices, and reporting on leaked figures suggests they are not cheap. Second, the headline demos - running very large language models locally - depend entirely on the memory tier: the 128GB flagship can hold models that the 32GB mainstream configuration physically cannot, and in both cases the LPDDR5X is soldered, so the capacity bought at checkout is the capacity the machine will always have. NVIDIA's chip story is entering the PC mainstream; what buyers get for their money is the part still to be presented.
The October 7 event is expected to fill in final performance figures and pricing across the lineup, with first units shipping through October.
Comments (0)
Log in to join the discussion
Log InNo comments yet