NVIDIA RTX Spark laptops went on sale on 16 October from ASUS, Dell, HP, Lenovo, Microsoft and MSI, with prices starting at $2,599 in the US. They are slim Windows machines built around one NVIDIA chip that combines a Blackwell RTX GPU, an Arm-based Grace CPU and up to 128GB of shared memory. That memory is the whole point: it lets a laptop load AI models that would not fit on a normal gaming GPU. NVIDIA calls this a global launch, but I could not find an official India price or sale date for these models as of 17 October.
What RTX Spark actually is
RTX Spark is a single “superchip” rather than a separate CPU and graphics card. NVIDIA pairs a Blackwell RTX GPU with its Grace CPU and says MediaTek helped design the CPU side for power efficiency. ASUS and Microsoft list the top configuration at up to 6,144 GPU cores, up to 20 CPU cores and up to 128GB of LPDDR5X unified memory. NVIDIA rates it at up to 1 petaflop of FP4 AI performance.
Because the CPU and GPU share one pool of memory, a big model does not have to squeeze into 8GB or 16GB of VRAM. Apple has done the same thing on Macs for years, which is why a Mac mini M6 or Mac Studio is popular with people who run models locally. RTX Spark brings that design to Windows, along with CUDA, which most AI tools are still built around.
There are two catches. First, the GPU cannot use all of that memory. Microsoft says the amount the GPU can address depends on the configuration and workload, and is less than the total. Second, the memory is soldered on. ASUS lists it as on-board LPDDR5X, so the configuration you buy is the one you keep.
RTX Spark laptop models at launch
NVIDIA lists nine launch models. The ASUS and Microsoft announcements include US starting prices; the other launch materials checked did not list one.
| Laptop | Key details from the maker | US starting price |
|---|---|---|
| ASUS ProArt P14 | 14-inch 3K OLED, 24GB to 128GB memory, 90Wh battery | $2,599.99 |
| ASUS ProArt P16 | 16-inch OLED up to 4K, 24GB to 128GB memory, 99.9Wh battery | $2,799.99 |
| Microsoft Surface Laptop Ultra | 15-inch PixelSense Ultra touchscreen, up to 128GB memory, removable storage | $2,599 |
| Dell XPS 16 | RTX Spark configuration | Not stated in launch announcement |
| HP OmniBook Ultra and OmniBook X 14 | RTX Spark configurations | Not stated in launch announcement |
| Lenovo Yoga Pro 9n | RTX Spark configuration | Not stated in launch announcement |
| MSI Prestige N16 Flip AI+ and Prestige N16 | RTX Spark configurations | Not stated in launch announcement |
Look past the “up to 128GB” headline. ASUS’s entry configurations start at 24GB. Microsoft’s $2,599.99 Surface also starts at 24GB, with an 18-core CPU and 5,120-core GPU; its 20-core CPU and 6,144-core GPU option starts at $3,699.99. If local AI is why you are buying one, compare the price of the 64GB or 128GB version rather than assuming the base model will handle the same workloads.
What it does for local AI agents
NVIDIA is selling RTX Spark as a computer for running AI agents on your desk instead of in the cloud. It says guided installers can set up GitHub Copilot, Hermes, OpenClaw, Perplexity, Pi and WorkBuddy without the command line. It also points to its own Nemotron 3.5 Bolt 74B-A9B model, which is tuned for tool calling and multi-step tasks, on supported configurations. Microsoft says Surface Laptop Ultra can run AI models with more than 120 billion parameters locally.
The safety pieces matter as much as the speed. NVIDIA’s OpenShell runtime and Microsoft Execution Containers are meant to limit what an agent can touch on your PC. If you plan to let an agent work on real files, check those permission controls before you check benchmarks.
RTX Spark also works with NVIDIA PAIR. One of these laptops can serve AI requests to other Windows, Linux and macOS machines on your home network, which suits a household or small studio with one powerful machine and several ordinary ones.
Creators and gamers: what NVIDIA promises
For video and 3D work, NVIDIA claims 8K 4:2:2 editing and 3D scenes of up to 90GB, even on battery. Photoshop and Premiere get native RTX Spark support through beta versions at launch, with Substance 3D coming later. For games, NVIDIA claims AAA titles at 1440p and over 100 frames per second, and it lists Cyberpunk 2077, Fortnite, Marvel Rivals, World of Warcraft and NBA 2K27 among the launch titles. Its GeForce team also lists anti-cheat games such as Counter-Strike 2 and Dota 2 as playable from day one.
These are NVIDIA’s numbers, not independent tests. The Arm CPU is the part to watch. NVIDIA and Microsoft say the first PyTorch CUDA build for Windows on Arm is ready, and that 97% of the top 15,000 Python packages are supported. That is a strong start, but if your work depends on an older plugin, a niche driver or a game with strict anti-cheat, check it before you buy.
RTX Spark laptops in India: what we know
NVIDIA says RTX Spark laptops are “available globally” through brand websites and major retailers, and that pricing and configurations vary by region. As of 17 October, I could not find an official India price or sale date for the listed models. The ASUS launch release only covers the US, with sales online at the ASUS Store and Best Buy first and Best Buy and Micro Center stores to follow.
Until there is an official Indian listing, be careful with “expected price” figures. Do not convert the US sticker price directly into rupees; an Indian retail price will include local taxes and may vary by configuration. Indian warranty and service matter too, since the memory and chip are soldered to the board. If you buy one abroad, check whether the brand covers it with an international warranty in India.
If you only need more memory for local models on an existing desktop, an upgrade may be the cheaper route. Our guide to RAM prices in India covers what desktop memory costs right now.
Should you buy an RTX Spark laptop?
Worth a look if you run local models or AI agents every day, edit heavy video, or want one portable Windows machine for CUDA work and gaming. For those uses, the 64GB or 128GB versions are the ones that justify the price.
Skip it, or wait, if you mainly game on Windows, depend on an old plug-in or driver without confirmed Windows on Arm support, or would end up buying the base 24GB model. A regular RTX gaming laptop will often be cheaper and more predictable for those jobs.
RTX Spark puts unusually large shared memory and CUDA in a portable Windows machine. That combination is promising for local AI, but independent tests of battery life, fan noise and app compatibility matter more than launch-day claims. I would wait for those reviews and an Indian price before buying.
FAQ
Is RTX Spark an Arm laptop?
Yes. The CPU side is NVIDIA’s Grace CPU, which is Arm-based, so these laptops run Windows on Arm. NVIDIA and Microsoft have prepared native developer tools, including a PyTorch CUDA build for Windows on Arm.
Can I upgrade the memory later?
No. The unified memory is on-board LPDDR5X, so pick the capacity you need at purchase. Surface Laptop Ultra does have user-removable storage, according to Microsoft.
How much does an RTX Spark laptop cost?
NVIDIA says prices start at $2,599 in the US. ASUS lists the ProArt P14 from $2,599.99 and the P16 from $2,799.99, and Microsoft lists Surface Laptop Ultra from $2,599. No Indian price has been announced yet.
Is RTX Spark the same as DGX Spark?
No. DGX Spark is NVIDIA’s small desktop AI computer for developers. RTX Spark is the chip platform for consumer Windows laptops and compact PCs, with RTX graphics and gaming support built in.
