Google DeepMind CEO Calls DeepSeek’s AI Cost Claims “Exaggerated”

Chinese AI startup DeepSeek’s claim of developing a powerful AI model at a much lower cost than major tech giants like Alphabet’s (GOOGL) Google, Meta (META), and Microsoft (MSFT) had sent shockwaves through the industry. However, Google DeepMind CEO Demis Hassabis recently called the buzz around DeepSeek’s low-cost model “exaggerated.”

He pointed out that DeepSeek’s claim of training its AI model for just $5.6 million, far less than the billions spent by other AI companies, has sparked skepticism. He suggested that this figure likely accounts only for the final training phase, overlooking the broader investment.

Moreover, he pointed out that the model is built on existing AI techniques without introducing any groundbreaking innovations. Also, Hassabis highlighted that Google’s Gemini 2.0 Flash models demonstrate superior efficiency compared to DeepSeek’s model.

DeepSeek’s Cost Claims Invite Regulatory Probes

DeepSeek’s cost claims have drawn scrutiny from researchers and regulatory authorities. Investigations are ongoing to determine if DeepSeek bypassed U.S. chip export restrictions by acquiring older Nvidia chips through Singapore.

Also, concerns have been raised about potential data theft, with allegations that DeepSeek may have used data from OpenAI’s models for training purposes.

It is worth noting that DeepSeek’s claims of a cost-effective AI model did little to deter major tech companies like Google, Meta, Microsoft, and Amazon (AMZN) from investing heavily in AI development. These companies disclosed plans to spend a combined $320 billion in 2025 on capital expenditures, a significant portion of which will be allocated to AI infrastructure development.

Which Is the Best AI Stock to Buy?

For investors interested in AI stocks, we have rounded up the best stocks that analysts are bullish about using the TipRanks Stocks Comparison tool.

Disclosure