Nvidia has finally completed its acquisition of Run:ai, which has been in the works since April when the GPU giant announced it wanted to purchase the AI software company. Run:ai offers GPU orchestration software to organize groups of GPUs more efficiently and has been a close partner with Nvidia since 2020; in fact, the only GPUs Run:ai currently supports are those made by Nvidia. The company was founded in 2018, around the same time Nvidia started dipping its toes into artificial intelligence with technologies like deep learning supersampling. The company says it will open-source its software.
Neither company has confirmed the deal's worth, but Nvidia reportedly spent around $700 million to acquire Run:ai.
Given that Nvidia has been so crucial for the AI industry with respect to hardware, it’s not surprising that the green giant is also emphasizing software, which represents vertical integration for Nvidia’s business. However, the GPU company already has a pretty good grip on software, thanks to its closed-source CUDA software, which has been around since 2007. CUDA has been a major advantage for Nvidia thanks to its entrenchment in the industry, and open-source competitors like AMD’s ROCm have struggled to find an opening.
The Run:ai acquisition could have increased the size of Nvidia’s walled garden even further, but Run:ai founders Omri Geller and Ronen Dar say that won’t be happening.
“True to our open-platform philosophy, as part of NVIDIA, we will keep empowering AI teams with the freedom to choose the tools, platforms, and frameworks that best suit their needs,” a press release written by the duo reads. “We will continue to strengthen our partnerships and work alongside the ecosystem to deliver a wide variety of AI solutions and platform choices.”
The two say they also plan on making Run:ai’s software open source, but no hard details were given with respect to what that would actually look like and when it would happen.
Why Nvidia would make Run:ai open source only after it completed the acquisition is unclear. It might have been a condition set by Run:ai itself since Geller and Dar cite their “open-platform philosophy” in their joint statement. Open-sourcing the software might also have helped convince regulators to allow the acquisition to go through, as Nvidia faced heavy scrutiny in both the US and the EU.
However, there might be a caveat to open-source Run:ai software: It isn’t clear whether Nvidia will offer expanded services with open source Run:ai software.
No matter how it goes, however, it seems Nvidia has expanded its influence in AI even further. That could pressure AMD and Intel, the two main rivals of Nvidia in the GPU space, to respond in some way, perhaps with their own acquisitions or partnerships.
