DeepSeek: A whale of a story of frugal innovation rattles tech world

OpenAI was founded as a lab in 2015 to take on the perceived dominance of Google in AI. Its meteoric rise over the last three years however made its dominance unchallengeable. But, Chinese LLM maker DeepSeek’s launch of its reasoning model R1 last week has challenged not just the dominance of OpenAI but has also questioned America’s supremacy in the global AI race.

DeepSeek with its logo of a majestic blue whale led to a massive sell off in the stock markets and debunked the theory that one needs to pump in billions to build and train AI models. The advent of DeepSeek has given hope to countries like India and its tech ecosystem to build their own AI models since the Chinese trailblazer has proven how models can be created with very little resources and thrifty innovation — something that India is very well capable of given its reputation for “jugaad”. ET’s Himanshi Lohchab, Annapurna Roy and Subhrojit Mallick uncover this whale of a story.

What Happened: India Impact DeepSeek’s rise led to clarion calls for building sovereign AI in India since the notion that it takes signifi cant investments and a lot of compute to build own language models has been debunked, questioning billions being poured by the likes of OpenAI, Nvidia, Microsoft and Google into GenAI.

It is said to have foiled US government’s attempts to restrict the march of other nations in what is being dubbed as the global AI race. Earlier this month, America has inforced export control on AI language models and GPUs for most nations including India and would have created an unlevel playing field especially for the developing nations.