The Spark is positioned as one of the world’s smallest AI supercomputers, bringing NVIDIA’s full AI stack into a compact, desktop-sized system. It combines NVIDIA’s Blackwell GPUs, Grace CPUs, networking, CUDA-X libraries, and the broader AI software stack, and is aimed at developers working on agentic and physical AI applications.
The system delivers 1 petaflop of AI performance with 128GB of unified memory in a small form factor. It is designed to help developers in India run inference on AI models with up to 200 billion parameters and fine-tune models of up to 70 billion parameters locally.
It also enables users to build AI agents and operate advanced AI software entirely on-premises, without relying on external cloud infrastructure.
In addition to the Spark, Netweb Technologies India Ltd has introduced Tyrone AI supercomputing systems built on NVIDIA’s Grace Blackwell platforms and manufactured in India. These systems are based on the NVIDIA GB200 NVL4 architecture, which integrates four NVIDIA Blackwell GPUs connected via NVLink and two NVIDIA Grace CPUs linked through NVLink-C2C.
The new systems are compatible with liquid-cooled NVIDIA MGX modular servers and are designed for high-performance workloads, including scientific computing, AI model training and inference. According to the company, they can deliver up to twice the performance of the previous generation in certain workloads.
The Tyrone Camarero GB200 AI system also incorporates technologies aimed at supporting large-scale AI training and real-time inference for large language models of up to 10 trillion parameters. These include the NVIDIA GB200 NVL4 chip, a second-generation Transformer Engine, fifth-generation NVLink, a RAS Engine, confidential computing features for secure AI processing, and a Decompression Engine.Netweb will showcase the Tyrone AI product range in New Delhi from February 16 to 20, 2026. The lineup will include the MGX-based GB200 liquid-cooled system and the Tyrone Camarero Spark, spanning applications from edge and personal AI computing to advanced data centre workloads, underscoring its India-based manufacturing push.
Sensex, Nifty today: Catch all the LIVE stock market action here
(Disclaimer: Recommendations, suggestions, views, and opinions given by the experts are their own. These do not represent the views of The Economic Times)