
Indian multinational corporation Infosys and their Vice President has stated that they are working towards creating a layer of software that decreases dependency on graphic processing units (GPUs) for AI tasks. This specific issue is very relevant today considering the growing demand for chips globally. Due to the high demand, the supply has become restricted which has dramatically increased chip prices.
Satish HC, the Executive Vice President and Chief Delivery officer of Infosys mentioned this development at NASSCOM Technology and Leadership Forum 2025. He claimed, “-Not every workload needs a GPU. We are working on a software layer that ensures AI models can run efficiently on commodity hardware….” It was stated that the software bypasses the need for costly GPUs.
As stated by Satish, “One of the things that we are very proud of sharing is that we have already built four models for our business within Infosys,” the company has made designs for smaller language models where a singular model is capable of creating and processing texts while being able to understand specific workload related data. This innovative working model was introduced by the company on January 16.
At the same time, Infosys has been exploring the efficiency of AI together with Wipro, which has focused $1 billion towards AI, pursuing AI implemented solutions. We ensured AI was incorporated in our businesses before deploying it to worksites. Our AI powered systems have enabled us to create relevant optimizations in functions like HR, Legal, Manufacturing and others, claimed Wipro’s COO, Sanjeev Jain.
GPU Shortage
India has mostly relied on NVIDIA, AMD and others for GPU supplies due to the unprecedented demand created by the global AI surge and India’s increasing needs. However, to tackle this challenge, the government is now proactively building the domestic semiconductor competency to lower the foreign supplier dependence.
India is establishing specialized chip manufacturing plants to address the semiconductor mission alongside Taiwan and other foreign powers. Although these plants will initially center around logic and memory chips, within three to five years, the nation will be able to produce its first domestically made GPU.
Read More: Samsung Galaxy S25 Plus Review: The Perfect Middle Ground in the Flagship Lineup