This comes as DeepSeek launched a new chatbot and AI model last month, developed at a fraction of the cost spent by US competitors like DeepMind and OpenAI.
In an interview with Bloomberg TV, Hassabis, who oversees Google’s AI division, explained that the Chinese start-up "only reported the cost of the final training round, which is just a small part of the total cost."
Subscribe today for free
The Chinese startup claimed to have spent $5.6 million on computing costs to train its model using older Nvidia chips.
However, numerous industry experts have raised doubts about these claims US authorities have launched an investigation to determine if DeepSeek bypassed a chip ban by making purchases through Singapore.
Meanwhile, Hassabis also dismissed the idea that DeepSeek has changed the economics of AI development.
He said: “We don’t see any new silver bullet technologies. DeepSeek is not an outlier on the efficiency curve.”
Instead, he pointed out that Google’s AI model, Gemini, is more efficient than DeepSeek in training-to-performance and cost-to-performance ratios but claimed the technology giant “doesn't talk about it very much.”
RELATED STORIES
Behind the DeepSeek hype: Costs, safety risks & censorship explained
DeepSeek failed all safety tests, responding to harmful prompts, Cisco data reveals
DeepSeek: Safety questions arise amid Australia ban & EU compliance uncertainty
