“Our market investigation confirmed to us that other software options compatible with Nvidia’s hardware will remain available in the market,” Teresa Ribera, the EU’s new antitrust chief, said in the statement.
The Commission had the option to either approve the deal, with or without conditions, or initiate a more in-depth four-month investigation if significant concerns arise.
Antitrust regulators in both Europe and the United States have increasingly raised concerns about “killer acquisitions,” where large corporations acquire startups only to shut them down and stifle competition.
In September, the U.S. Department of Justice (DOJ) issued subpoenas to Nvidia and other companies as part of an investigation into potential antitrust violations. This investigation includes Nvidia’s acquisition of Run:ai, which develops an operating system for AI processors. Concerns have been raised that integrating Run:ai’s system with Nvidia’s other products could make it harder for customers to switch to competing chips. Investigators are also examining whether Nvidia incentivizes customers with special pricing or supplies to exclusively use its technology or purchase complete systems.
Although the deal does not meet the EU turnover threshold that would typically mandate Nvidia to request EU approval, the Italian competition authority notified the European Commission about the case. The EU accepted the referral and warned about potential competition risks posed by the deal.
Run:ai raised $75 million in a Series C funding round in March 2022, led by Tiger Global Management and Insight Partners, who also led the Series B round. Additional participation came from existing investors TLV Partners and S Capital VC, bringing Run:ai’s total funding to $118 million.
Founded in 2018 by Omri Geller (CEO) and Dr. Ronen Dar (CTO), Run:ai specializes in orchestration and virtualization software designed for the unique requirements of AI workloads on GPUs and similar chipsets. Its Kubernetes-based container platform for AI clouds optimizes GPU utilization by dynamically allocating resources—ranging from fractions of GPUs to multiple GPUs across nodes—ensuring efficient and scalable performance.
Reuters contributed to this report.
