India is preparing to compete in the global artificial intelligence (AI) race and plans to launch its own generative AI model in 2025, according to Union IT Minister Ashwini Vaishnaw. He made this announcement at the Utkarsh Odisha Conclave.
As reported by The Economic Times of India, the country has secured 18,693 GPUs, including 12,896 high-performance Nvidia H100 chips. Additionally, India is expecting around $20 billion in foreign investments to build data centers over the next three years.
The minister also gave a timeline for India’s AI model, which will be designed to support multiple Indian languages and cultures. He stated, “We believe at least six major developers can create AI models within six to eight months, or even in four to six months under a more optimistic estimate.”
India’s AI push comes shortly after the release of DeepSeek R1, an open-source AI model that competes with OpenAI’s leading models but was developed at a much lower cost. This has challenged previous assumptions that AI development always requires massive computing power.
Meanwhile, in response to DeepSeek’s success, former US President Donald Trump is considering stricter export rules on high-performance AI chips made by Nvidia. The US has already restricted Nvidia from selling its H100 chips to China since 2022. In 2023, additional bans were placed on semiconductor sales, including weaker AI chips like Nvidia’s A800 and H800, which were initially designed to bypass earlier restrictions.
Trump has pledged to make the US a leader in AI and semiconductor technology. He recently announced “Project Stargate,” a $500 billion initiative involving OpenAI, Oracle, and SoftBank to build AI infrastructure in the US.
However, some experts warn that tighter regulations on US companies could weaken the country’s AI leadership. They argue that such measures may allow smaller and more flexible competitors to take the lead in AI development.
 race and plans to launch its own generative AI model in 2025,){kind=link}