Ziroh Labs, a US-based AI startup, in collaboration with the Indian Institute of Technology, Madras (IIT Madras), has unveiled Kompact AI, a system designed to run large AI models using CPUs instead of advanced computing chips.
This initiative aims to address India’s compute accessibility challenges by developing efficient AI solutions for CPU and edge device inferencing, the institute said in a statement.
Go deeper with GlobalData
Kompact AI is part of India’s AI mission, “AI for All”, and is developed by Indian scientists and engineers.
The platform allows foundational AI models to be built and served on CPUs, eliminating the need for costly, hard-to-obtain GPUs.
In a joint statement, Ziroh Labs co-founders Hrishikesh Dewan and Igneta DSouza said: “AI has the potential to transform every aspect of our lives, from improving healthcare outcomes to enhancing education and solving agricultural issues.
“However, for AI to benefit everyone, it must be accessible, affordable, and adaptable to real-world conditions. Kompact AI is designed for such needs and operates efficiently on CPUs, optimised for task-specific applications. This ensures that communities and enterprises with limited infrastructure can leverage the power of AI to create practical, impactful solutions.”
Ziroh Labs has optimised 17 AI models, including DeepSeek, Qwen, and Llama, to run efficiently on CPUs.
These models have been benchmarked with IIT Madras, assessing both quantitative performance and qualitative accuracy.
Kompact AI does not require constant internet connectivity, making it suitable for use in remote areas with limited or no access to the internet.
The startup’s technology has been tested by US chipmakers Intel and Advanced Micro Devices, according to Bloomberg.
The collaboration also plans to establish a centre of AI research to further develop practical AI solutions.
