On March 11th, Huang Renxun, the CEO of Nvidia, stated in an interview with John Shoven, the Honorary Professor of Economics at Stanford University, that even if competitors' AI chips were free, it would be difficult to shake Nvidia's position.
Huang further explained that, in terms of the total cost of ownership (TCO) in AI data centers, the pricing of Nvidia's GPUs is not the most important factor. Nvidia has undeniably achieved impressive accomplishments in driving the development of the AI industry. Recently, the company became the world's third-largest by market capitalization, largely due to its AI-accelerating GPUs. However, Huang's remarks are likely to be controversial as he holds a dismissive stance towards many competitors, such as AMD, Intel, and a range of companies with ASICs and other custom AI chips.
In a video excerpt from the interview released by Stanford Institute for Economic Policy Research (SIEPR), Charles R. Schwab Professor of Economics and former Trione Director at SIEPR, John Shoven, asked, "Is it possible that someone could come along and say, 'Well, we've got a chip that's just as good as Nvidia's, but it's a lot cheaper?' Is that a threat?"
Huang responded, "We compete with everyone on the planet, and in some cases, even Nvidia's customers are competitors. Furthermore, Nvidia actively assists customers who are designing alternative AI processors, even sharing Nvidia's upcoming chip roadmap with them."
Huang emphasized that Nvidia operates in a "fully open" manner and collaborates with almost everyone in the industry. While chip buyers consider the price, data center operators consider operational costs. However, the company recognizes the importance of TCO, which encompasses deployment time, performance, utilization, and flexibility. Even if competitors' chips were free, they would still not be cheap enough.
Huang concluded his response to Shoven by emphasizing Nvidia's goal of maintaining an unparalleled TCO leadership position.
In the interview video, Huang outlined Nvidia's currently unassailable unique selling points. He mentioned that while one could manufacture an ASIC chip that excels in a specific algorithm, Nvidia's GPUs are programmable. Additionally, Nvidia's platform has become a significant standard for every major cloud computing company. As a result, typical data centers across various industries, from financial services to manufacturing, are inclined to choose Nvidia hardware.
ALSO READ:Latest Ranking of Global IC Design Companies: NVIDIA Takes a Commanding Lead!
