In response to news that Nvidia has received a subpoena from the US Department of Justice related to potential antitrust investigations, Nvidia responded by saying, “We have made inquiries to the US Department of Justice but have not received a subpoena yet. Nvidia wins by strength. We are happy to answer any questions regulators may have about our business.”
The market value evaporated by 400 billion dollars in a week.
The issuance of a subpoena by the US Department of Justice would mean an escalation of the antitrust investigation against the company. According to previous news, the US Department of Justice has been asking technology companies about Nvidia’s business practices in recent weeks, including problems with the company’s hardware bundling sales behavior.
In addition to bundling sales, the US Department of Justice is also investigating Nvidia’s investment and acquisition transactions, including a $700 million transaction in which Nvidia acquired Israeli company Run:AI in April this year.
Since Nvidia released its latest quarterly financial report last week, its stock price has fallen continuously in recent days. The stock price has fallen by more than 15% in the past five trading days, and its market value has evaporated by more than $400 billion. On September 3 alone, the company’s stock price fell by nearly 10%, and the net worth of Nvidia founder and CEO Jensen Huang lost nearly $10 billion.
Although the US Department of Justice has not yet filed an antitrust lawsuit against Nvidia, the potential escalation of antitrust investigations also reflects Nvidia’s growing control in the rapidly expanding artificial intelligence computing market. According to industry estimates, Nvidia currently accounts for more than 80% of the market share of AI chips for data centers.
Antitrust agencies believe that Nvidia tries to “bind” users through a powerful software ecosystem platform, making it more difficult for users to switch to other AI chip suppliers.
Earlier investigations by the French antitrust regulator believed that there is a risk of abuse of dominant position by AI chip suppliers. The agency particularly emphasized that the excessive dependence of the industry on Nvidia’s CUDA chip programming software is worrying, and also mentioned being uneasy about Nvidia’s investment in CoreWeave and other cloud service providers focused on artificial intelligence.
Benoit Coeure, chairman of the French Competition Authority, said that if the investigation yields results, it means that Nvidia will be sued by the French antitrust agency. Companies that violate French antitrust rules may be fined up to 10% of their global annual turnover.
In a regulatory filing last year, Nvidia said that regulatory agencies in the European Union, China, and France have asked it to provide relevant information about GPU graphics cards.
Benefiting from its dominant position in the field of AI chips for data centers, Nvidia’s performance has grown significantly in recent years, and its revenue has achieved triple-digit growth for four consecutive quarters. In the second quarter of this year, Nvidia’s sales exceeded $30 billion, and its profit more than doubled.
Dominance was established ten years ago.
In the era of generative AI, CUDA software, as a tool for providing developers to run artificial intelligence programs on Nvidia chips, has become an important part of Nvidia’s corporate strategy. Nvidia allows developers to create and deploy generative AI assistants in systems with installed CUDA GPUs.
As Nvidia’s sales of AI chips continue to increase, the company has also upgraded its new enterprise software subscription and marketed its network products as an important supplement, thus helping enterprises give full play to the effectiveness of chips. And Nvidia’s chips are usually pre-installed in entire server racks designed by Nvidia, so that Nvidia has transformed from a simple parts supplier to an integrated system supplier, which can generate a greater premium effect.
Regarding Nvidia’s business model, Sheng Linghai, an analyst at research institute Gartner, told Yicai reporter: “Essentially, Nvidia’s software business is more profitable, which is closely related to the software and service-centered structure of the US technology industry.”
Strictly speaking, Nvidia officially does not support the compatibility of the CUDA software platform with other chips. However, since CUDA has been an open source platform since its launch, there are also manufacturers including Chinese chip companies that claim that their chips are compatible with CUDA. But analysts say that it is still unknown how much actual efficiency these chips can achieve.
Nvidia’s CUDA currently does not charge licensing fees, but in fact software can generate higher profits. But on the contrary, the fact that CUDA was completely open source without charging fees from the very beginning is the most crucial factor for its success.
Judging from the financial report, Nvidia’s gross margin in the second quarter dropped from 78.4% in the previous quarter to 75.1%, but it is still higher than 70.1% in the same period last year. The company expects that the annual gross margin will remain around 70%, but it is lower than the analyst’s expectation of 76.4%.
On the other hand, although global antitrust regulatory agencies have been keeping an eye on Nvidia, so far there is no very reasonable evidence to prove that it has violated antitrust laws.
Reproduction Notes：InvestFancy » Ntitrust Agencies are ‘Besieging’ Nvidia