BENGALURU: With DeepSeek, China has emerged as the new powerhouse in AI foundation models. From being the factory of the world, it looks to now be emerging as the world’s LLM (large language model) factory – developing frontier AI models at a fraction of the cost in the West.
US President Donald Trump was quoted as saying that DeepSeek “should be a wake-up call for our industries that we need to be laser-focused on competing to win”.
Poll
Do You Think India Can Compete with China in the AI Race?
DeepSeek even won the appreciation of OpenAI chief executive Sam Altman, who said in a post on X that it was “invigorating to have a new competitor,” and described DeepSeek’s R1 as “an impressive model, particularly around what they’re able to deliver for the price.”
How should India respond to this technological advance? Over the past year, the country has split into two factions – one that wants to build indigenous LLMs from scratch, and the other that wants to build small language models with fewer parameters and which focus on specific applications. Homegrown AI startup Sarvam AI’s platform was trained on 2 billion parameters, with an emphasis on Indian languages. R1 was trained on 671 billion parameters, and does not focus on any specific use case.
Viral Shah, co-inventor of Julia programming language and who was part of the early UIDAI team that built Aadhaar, says India has capital, talent, and competence in AI. He notes that marquee global VC firms have offices in India, and a significant chunk of Nvidia and Intel chips are designed in India. “It will take time, the entire ecosystem needs to band together to get our fundamentals right by bettering ease of business, improving higher education, and increasing R&D spends on deeptech,” he says.
Shah adds that Indian entrepreneurs, business houses, and venture capitalists must invest time and money into hundreds of new ideas for the country to have a standing in AI. “Frugality cannot be an excuse to miss out on being the best in the world,” he says.
Arun Chandrasekaran, distinguished VP analyst on AI at Gartner, says India needs to invest more in core research to build some capable models in the global AI race. “It needs to spend a lot more on core infrastructure, state-of-the-art data centres, energy and cooling technologies, compute and chips. Beyond that, also in research at our universities. Getting the smartest people together in a room and equipping them with the right infra and compute may just be the magic we need,” he says.
Domestic electronics manufacturer Indkal Technologies founder and CEO Anand Dubey notes that private companies in China, historically, have always focused their time and investments on commoditising hardware.