Huaren Huang: Competitors' free AI chips can't do anything with Nvidia GPU
"Free is not cheap enough" to compete with the "green team" in the AI field.
Recently, NVIDIA CEO Jensen Huang claimed: "NVIDIA's GPUs are so good that even if a competitor's chips are free, they are not cheap enough. In terms of the total cost of ownership (TCO) for AI data centers, the pricing of NVIDIA GPUs does not matter."
It is undeniable that NVIDIA's achievements in driving the booming artificial intelligence industry are impressive. The company has recently become the world's third most valuable (by market capitalization) company, largely thanks to its AI-accelerating GPUs. However, Huang's comments will certainly spark controversy, as he dismisses a host of competitors, such as AMD, Intel, and a range of competitors with ASICs and other types of custom AI chips.
Former SIEPR Trione Director and Stanford University Charles R. Schwab Professor of Economics Emeritus John Shoven asked: "You have designed the most advanced chip, is there a possibility of facing competition that claims to be good enough, not as good as NVIDIA, but much cheaper? Is this a threat?"
Huang said: "We face more competition than anyone on the planet," he told Shoven, in some cases, even NVIDIA's customers are its competitors. In addition, Huang pointed out that NVIDIA actively helps customers who are designing alternative AI processors, even revealing the upcoming NVIDIA chips on the roadmap to them.
Advertisement
This does sound like an unusual way of doing business. Huang claims that NVIDIA is completely open while collaborating with almost everyone else in the industry, which may be harder to believe. Just recently, there have been strong accusations against NVIDIA, with reports that customers are afraid to talk to competitors' GPU/AI accelerator manufacturers for fear of order delays. An industry alliance has also been formed to try to overturn the company's dominant CUDA programming model.
Back at the Stanford University interview, Huang outlined NVIDIA's current unique selling points, stating: "While you can make a chip that is good at a specific algorithm, NVIDIA's GPUs are programmable. Moreover, the NVIDIA platform is a great standard... in every cloud computing company. Therefore, a typical data center that wants to support a wide range of customers from financial services to manufacturing will be attracted by NVIDIA hardware."
Free is not cheap enough.Jensen Huang also tried to provide a contrast, stating that people who buy and sell chips consider the price of the chips, while those who operate data centers consider the operational costs. Of course, the company is well aware of the total cost of ownership (TCO), which essentially means that the benefits claimed by NVIDIA, such as deployment time, performance, utilization, and flexibility, are so significant that "even if a competitor's chips were free, they wouldn't be cheap enough," said Huang.
In concluding his response to Shoven, the CEO emphasized that maintaining a leading position with an unparalleled TCO is NVIDIA's goal. Huang reminded the summit attendees that what NVIDIA does requires a lot of effort and innovation—it's not about luck—nothing is taken for granted.
Naturally, the views of NVIDIA's competitors differ from Huang's statements, but for now, NVIDIA is the undisputed king of the AI world.
NVIDIA is described by competitors as a "cartel" controlling GPU supply.
According to statements from Groq executive and former AMD executive Scott Herkelman, NVIDIA, the most important driver of the global AI craze, is facing accusations of a "GPU cartel" and controlling the supply of data center markets.
In an interview, Groq CEO Jonathan Ross claimed that some of NVIDIA's data center customers are even afraid to meet with competitors in the AI chip manufacturing industry because they fear retaliation from NVIDIA, which could delay the shipment of ordered GPUs. Despite NVIDIA's claim that it is trying to fairly allocate supplies during the global shortage.
"This happens more than you think. NVIDIA does this with DC customers, OEMs, AIBs, media, and distributors. They learned from GPP not to put it in writing. They only ship after customers place orders," said former Senior Vice President and General Manager of AMD Radeon, Scott Herkelman, in response to accusations on X (Twitter): "They are the GPU cartel organization, controlling all supplies."
These comments refer to NVIDIA's GeForce Partner Program (GPP) in 2018, which was abandoned due to backlash over its exclusive requirements. Herkelman implies that NVIDIA continues similar practices but avoids written agreements.
The report also suggests that large tech companies like Microsoft, Google, and Amazon are developing their own AI accelerators but downplay their position as competitors to NVIDIA. This further indicates that NVIDIA is perceived as controlling access to key technologies for AI development.NVIDIA currently provides computational support for approximately 80% of global artificial intelligence development, wielding an incredible influence over strategic technologies. The allegations from Groq and Herkelman suggest that the company is willing to actively leverage this market position to protect its dominance.
NVIDIA has not yet officially responded to the latest accusations. However, as regulatory scrutiny of the market power of tech giants becomes increasingly stringent, these reports intensify speculation about anti-competitive practices, potentially subjecting NVIDIA to regulatory pressure. The company will likely need to transparently address whether its supply distribution favors certain customers based on relationships with competitor chip manufacturers.
*Disclaimer: This article is the original creation of the author. The content of the article represents their personal views, and our reposting is solely for sharing and discussion, not an endorsement or agreement. If you have any objections, please contact the support team.
"Free is not cheap enough" to compete with the "green team" in the AI field.
Recently, NVIDIA CEO Jensen Huang claimed: "NVIDIA's GPUs are so good that even if a competitor's chips are free, they are not cheap enough. In terms of the total cost of ownership (TCO) for AI data centers, the pricing of NVIDIA GPUs does not matter."
It is undeniable that NVIDIA's achievements in driving the booming artificial intelligence industry are impressive. The company has recently become the world's third most valuable (by market capitalization) company, largely thanks to its AI-accelerating GPUs. However, Huang's comments will certainly spark controversy, as he dismisses a host of competitors, such as AMD, Intel, and a range of competitors with ASICs and other types of custom AI chips.
Former SIEPR Trione Director and Stanford University Charles R. Schwab Professor of Economics Emeritus John Shoven asked: "You have designed the most advanced chip, is there a possibility of facing competition that claims to be good enough, not as good as NVIDIA, but much cheaper? Is this a threat?"
Huang said: "We face more competition than anyone on the planet," he told Shoven, in some cases, even NVIDIA's customers are its competitors. In addition, Huang pointed out that NVIDIA actively helps customers who are designing alternative AI processors, even revealing the upcoming NVIDIA chips on the roadmap to them.
Advertisement
This does sound like an unusual way of doing business. Huang claims that NVIDIA is completely open while collaborating with almost everyone else in the industry, which may be harder to believe. Just recently, there have been strong accusations against NVIDIA, with reports that customers are afraid to talk to competitors' GPU/AI accelerator manufacturers for fear of order delays. An industry alliance has also been formed to try to overturn the company's dominant CUDA programming model.
Back at the Stanford University interview, Huang outlined NVIDIA's current unique selling points, stating: "While you can make a chip that is good at a specific algorithm, NVIDIA's GPUs are programmable. Moreover, the NVIDIA platform is a great standard... in every cloud computing company. Therefore, a typical data center that wants to support a wide range of customers from financial services to manufacturing will be attracted by NVIDIA hardware."
Free is not cheap enough.Jensen Huang also tried to provide a contrast, stating that people who buy and sell chips consider the price of the chips, while those who operate data centers consider the operational costs. Of course, the company is well aware of the total cost of ownership (TCO), which essentially means that the benefits claimed by NVIDIA, such as deployment time, performance, utilization, and flexibility, are so significant that "even if a competitor's chips were free, they wouldn't be cheap enough," said Huang.
In concluding his response to Shoven, the CEO emphasized that maintaining a leading position with an unparalleled TCO is NVIDIA's goal. Huang reminded the summit attendees that what NVIDIA does requires a lot of effort and innovation—it's not about luck—nothing is taken for granted.
Naturally, the views of NVIDIA's competitors differ from Huang's statements, but for now, NVIDIA is the undisputed king of the AI world.
NVIDIA is described by competitors as a "cartel" controlling GPU supply.
According to statements from Groq executive and former AMD executive Scott Herkelman, NVIDIA, the most important driver of the global AI craze, is facing accusations of a "GPU cartel" and controlling the supply of data center markets.
In an interview, Groq CEO Jonathan Ross claimed that some of NVIDIA's data center customers are even afraid to meet with competitors in the AI chip manufacturing industry because they fear retaliation from NVIDIA, which could delay the shipment of ordered GPUs. Despite NVIDIA's claim that it is trying to fairly allocate supplies during the global shortage.
"This happens more than you think. NVIDIA does this with DC customers, OEMs, AIBs, media, and distributors. They learned from GPP not to put it in writing. They only ship after customers place orders," said former Senior Vice President and General Manager of AMD Radeon, Scott Herkelman, in response to accusations on X (Twitter): "They are the GPU cartel organization, controlling all supplies."
These comments refer to NVIDIA's GeForce Partner Program (GPP) in 2018, which was abandoned due to backlash over its exclusive requirements. Herkelman implies that NVIDIA continues similar practices but avoids written agreements.
The report also suggests that large tech companies like Microsoft, Google, and Amazon are developing their own AI accelerators but downplay their position as competitors to NVIDIA. This further indicates that NVIDIA is perceived as controlling access to key technologies for AI development.NVIDIA currently provides computational support for approximately 80% of global artificial intelligence development, wielding an incredible influence over strategic technologies. The allegations from Groq and Herkelman suggest that the company is willing to actively leverage this market position to protect its dominance.
NVIDIA has not yet officially responded to the latest accusations. However, as regulatory scrutiny of the market power of tech giants becomes increasingly stringent, these reports intensify speculation about anti-competitive practices, potentially subjecting NVIDIA to regulatory pressure. The company will likely need to transparently address whether its supply distribution favors certain customers based on relationships with competitor chip manufacturers.
*Disclaimer: This article is the original creation of the author. The content of the article represents their personal views, and our reposting is solely for sharing and discussion, not an endorsement or agreement. If you have any objections, please contact the support team.