Former Google CEO Believes NVIDIA is the Unmatchable Leader in AI Chips

As the undisputed leader in the field of AI chips, NVIDIA's charm is irresistible, even former senior executives of tech giant Google have been cheering for it.

Former Google CEO Eric Schmidt recently said that although he is not good at providing investment advice, he saw a clear opportunity in the stock market, which would NVIDIA.

In a video released this week, Schmidt said that large technology companies are planning to invest more and more in NVIDIA-based artificial intelligence data centers, with construction costs possibly reaching $300 billion. However, this video was later deleted.

I'm talking to the big companies, and the big companies are telling me they need $20 billion, $50 billion, $100 billion - very very hard, said Schmidt.

Schmidt said that a large part of this expenditure will go to NVIDIA, which produces dominant data center AI chips and has achieved more than 200% revenue growth for three consecutive quarters.

Google has also developed a chip called Tensor Processing Unit (TPU), which can compete with NVIDIA's processors, but it is still in the early stages.

If $300 billion is all going to Nvidia, you know what to do in the stock market, said Schmidt.

However, Schmidt did not disclose whether he holds NVIDIA's stock.

It is worth noting that the video was removed at Schmidt's request because he made inappropriate remarks about Google's office culture in the video, claiming that Google lost in the AI race because the company's employees were not working hard enough.

Schmidt served as Google's CEO and chairman from 2001 to 2011, then he handed over the CEO position to Google co-founder Larry Page, and he continued to serve as Google's executive chairman and technical advisor, and finally left Google in early 2020.

According to media reports, Schmidt founded the venture capital firm Innovation Endeavors in 2010 and currently holds about 147 million shares of Alphabet, worth about 24 billion US dollars. In addition, he is also a philanthropist and provides technical advice for several government committees.

It is worth noting that although the demand for NVIDIA chips by cloud computing companies and leading AI model developers has soared, Wall Street is currently considering a question, that is, whether NVIDIA's top customers are over-investing in AI infrastructure. NVIDIA will provide the latest demand information when it announces quarterly results on August 28th.

Schmidt said that NVIDIA will not be the only winner in artificial intelligence, but there are not many other obvious choices. He said that he currently believes that companies that can invest more money in NVIDIA chips and data centers will be technologically ahead of small competitors who cannot invest as much as they do.

At the moment, the gap between the frontier models - there are only three - and everyone else appears to be getting larger. Six months ago, I was convinced that the gap was getting smaller, so I invested lots of money in the little companies. Now I'm not so sure, said Schmidt.

Schmidt said that competitors will find it difficult to catch up with NVIDIA because many of the most important open-source tools used by artificial intelligence developers are based on the company's CUDA programming language.

He also pointed out that AMD's software to convert NVIDIA CUDA code and run it on its chips is currently not working.