Nvidia faces a U.S. Department of Justice antitrust investigation into complaints about its dominance in artificial intelligence chips. The DOJ contacted Nvidia competitors and AI chip startups while gathering information about alleged sales practices and acquisitions.
Currently, investigators are examining whether the company pressured customers to buy multiple products or favored customers that purchased its equipment. The inquiry does not establish wrongdoing. Instead, officials must determine whether Nvidia’s conduct unlawfully restricts competition.
DOJ Targets Nvidia’s AI Chip Business
At the moment, the DOJ is examining whether the company pressured cloud providers to buy products alongside its AI chips. Rival representatives alleged bundles could lower product prices. Another allegation says the company restricted chip supplies when customers declined additional products.
In addition, investigators asked whether the company charges higher prices for networking equipment when customers buy competing AI chips. Those allegations remain under investigation and the DOJ has not announced any findings.
Furthermore, the agency is examining Nvidia’s acquisitions of Run:ai and Bright Computing. Run:ai develops software that helps customers schedule workloads across data-center GPUs. The DOJ wants information about whether the company could use the acquisition to limit support for rival chips.
On a separate occasion, the European Commission examined the company’s Run:ai acquisition. Its decision found that alternatives to Run:ai would remain available. Therefore, the Commission found no ability to run competitors out of business.
Why Nvidia’s Market Power Matters
Currently, Nvidia controls more than 90% of the data-center GPU market. GPUs provide computing power for training and running AI systems. However, customers and rivals worry about dependence on one supplier.
On the other hand, CUDA creates another competitive concern. Nvidia developed the programming platform for its GPUs and developers widely use it for machine-learning applications. As a result, rival chipmakers argue that CUDA can make switching hardware harder because developers already use the company’s software ecosystem.
The DOJ is also examining sales practices involving chips and networking products. Cloud providers reportedly buy about half the AI chips the company makes. Although, offering discounts through bundling is not illegal on its own but tying a dominant product can raise antitrust concerns.
The Probe Could Reshape AI Competition
However, the investigation could last years before the DOJ decides whether to file a lawsuit. Antitrust investigations do not automatically lead to enforcement actions. Therefore, the process could end without making it to court.
For now, regulators are gathering evidence about Nvidia’s sales practices, acquisitions and competitive position. Their findings could influence how Nvidia structures product bundles and software relationships. Ultimately, the outcome could in turn affect how rival chipmakers compete for customers as demand for AI computing continues to grow.
