As AI grows in popularity, India is preparing to introduce its own LLM model as part of the IndiaAI Mission, according to IT Minister Ashwini Vaishnaw at the Utkarsh Odisha Conclave on Thursday. This comes just days after DeepSeek introduced its low-cost AI model R1, which outperformed OpenAI’s ChatGPT.
The Centre also stated that it has selected ten companies to provide high-end chips, 18,693 graphics processing units (GPUs), including the NVIDIA H100 and H200, as well as MI325 models, in order to develop the machine learning tools that will be used to develop the foundation model. Yotta, backed by the Hiranandani Group, Jio Platforms, Tata Communications, E2E Networks, CMS Computers, Ctrls Datacenters, Locuz Enterprise Solutions, NxtGen Datacenter, Orient Technologies, and Vensysco Technologies are among the companies that will provide GPUs.
Also Read: DeepSeek AI tops app store charts but is restricting new users amid malicious server attacks
For those unfamiliar, DeepSeek AI is trained on 2,500 GPUs, while ChatGPT is trained on 25,000. Now, India has stated that its LLM will be trained on 15,000 high-end GPUs, indicating that it will have a strong position in the AI race.
READ: OpenAI engineer’s viral post: Indian mom’s hilarious response to daughter’s tech achievement
“Over the last 1.5 years, our teams have been working closely with startups, researchers, professors etc. Today, we are calling for proposals to develop our own foundational model. The model will take care that the Indian context, languages, and culture are devoid of biases,” Vaishnaw added.
He also stated that the government is in talks with more than six developers to build the foundational model, which could take up to eight months. Furthermore, the reports state that the center will establish a common compute facility, and startups and researchers will be able to access computing power within the next few days.