NVIDIA launches RTX Spark laptops with 128GB unified memory in India
NVIDIA's new platform brings high-end AI performance and 8K video editing capabilities to slim laptops and compact desktops.
The short version
- The platform features up to 128GB of unified memory, allowing CPU and GPU to share the same memory pool.
- It delivers up to 1 petaflop of FP4 AI performance for local tasks like running AI agents.
- Supported hardware includes slim laptops from ASUS, Dell, HP, Lenovo, and MSI.
- The system supports 8K 4:2:2 video editing and rendering 3D scenes up to 90GB on battery power.
Why it matters
This shift moves high-end AI capabilities from massive server racks into portable workstations. It allows creators and developers to run complex models and high-resolution video workflows locally without relying on the cloud.
A new class of portable AI
NVIDIA recently showcased its new RTX Spark platform in India, marking a shift toward high-performance AI in compact hardware. The platform is designed to handle demanding AI models and creative workloads locally on slim Windows laptops and small desktops. This move positions these machines as portable AI workstations rather than standard thin-and-light laptops.
The launch event in New Delhi featured several major manufacturers, including ASUS, Dell, HP, Lenovo, and MSI. Each brand presented different form factors, ranging from the 17.8mm-thin Dell XPS 16 Creator Edition to the Lenovo Yoga Pro 9n. These devices are built to provide GPU-class AI performance in portable frames that users can carry easily.
By bringing these capabilities to the consumer market, NVIDIA aims to make local AI agents a primary use case for PC owners. Users can run agents like GitHub Copilot, Hermes, and Perplexity without needing complex command-line setups. The goal is to provide a seamless experience where AI tools are integrated directly into the hardware's capabilities.
The superchip architecture
At the heart of the RTX Spark platform is a specialized superchip that combines an NVIDIA Blackwell RTX GPU with an NVIDIA Grace CPU. This architecture is unique because it utilizes up to 128GB of shared unified memory. This means the CPU and GPU do not have to rely on separate memory pools, which can often create bottlenecks during heavy tasks.
MediaTek collaborated with NVIDIA on the custom CPU design to prioritize power efficiency and connectivity. This partnership ensures that the high-performance components can operate effectively within the power constraints of a laptop. The platform also integrates standard NVIDIA technologies like CUDA, DLSS, and TensorRT to support a wide range of applications.
The hardware is capable of delivering up to 1 petaflop of FP4 AI performance. This specific metric refers to the speed at which the system can process AI calculations using a specific data format. Such power allows the device to run large models, such as the Nemotron 3.5 Bolt 74B-A9B, which is designed for multi-step tasks and tool calling.
Creative and professional tools
The RTX Spark platform is specifically engineered for professional creative workflows that require significant memory and processing power. For example, the hardware is capable of processing 8K 4:2:2 video files and rendering 3D environments as large as 90GB. Notably, these demanding tasks can be performed while the device is running on battery power, which is a significant milestone for mobile workstations.
Software partners are already integrating support for this new hardware to enhance professional tools. Adobe plans to include TensorRT speedups for AI-driven functions within Photoshop, Premiere, and Substance 3D. Native RTX Spark support is already available in beta releases for Photoshop and Premiere, with Substance 3D support expected to arrive later.
Additional features include NVIDIA Broadcast 2.3, which offers background-noise removal for existing recordings and Studio Voice Mic Profiles. The platform also supports beta RTX Video features that can provide 4x frame generation and 4K super resolution. These tools are designed to give creators more control over their output quality and performance.
Market availability and pricing
Several manufacturers have already begun introducing RTX Spark-powered devices in the Indian market with varying price points. The Lenovo Yoga Pro 9n and Yoga 9n 2-in-1 are among the more accessible options, starting at Rs 2,99,399 and Rs 3,32,801, respectively. These models target creators and developers who need on-device AI performance in a portable form factor.
Higher-end options include the Dell XPS 16 Creator Edition, which starts at Rs 4,49,990, and the HP OmniBook Ultra, starting at Rs 5,49,999. For those seeking extreme performance, the ASUS ProArt GR1X is available starting at Rs 9,00,000. These prices reflect the premium nature of the unified memory and Blackwell GPU architecture.
Other notable entries include the ASUS ProArt P14 and P16, which start at Rs 3,19,990 and Rs 4,89,990, respectively. MSI also offers the EdgeMesa N AI+ desktop solution starting at Rs 3,89,999 plus GST. Gigabyte is expected to launch the AI TOP Aero in December with a starting price of Rs 7,35,499.
What to watch
- How many more professional creative apps will add native RTX Spark support?
- Whether the 128GB unified memory becomes the new standard for 'AI PC' marketing.
Try it
Check if your current laptop can run local AI models by looking for 'unified memory' specifications in its hardware details.
Sources and links
- ReportingMySmartPrice: the article we sourced
First reported by MySmartPrice. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.