Under antitrust pressure? After NVIDIA's $700 million acquisition, AI startup Run:ai plans to open source its software
Run:ai's software only runs on systems based on NVIDIA, and NVIDIA stated that the two companies have been closely cooperating since 2010. The U.S. Department of Justice and the European Union previously initiated investigations into whether NVIDIA's acquisition of Run:ai would harm competition. Just over a week ago, the European Commission determined that the acquisition would not hinder competition in the AI and GPU sectors, allowing the transaction to proceed
After overcoming regulatory hurdles, the Israeli artificial intelligence (AI) startup Run:ai, recently acquired by NVIDIA, is set to open source its software.
On Monday, December 30th, Eastern Time, Run:ai announced on its official website that its software, currently only running on NVIDIA systems, will be open sourced. This means that NVIDIA's competitors, such as AMD and Intel, will be able to access Run:ai's code and adapt it for computers using competing hardware.
Run:ai stated, "We are eager to build on our achievements to date, expand our excellent team, and broaden our product and market coverage. Open sourcing the software will allow it to extend across the entire AI ecosystem."
Run:ai's software helps manage and optimize AI hardware infrastructure, aiming to make it easier for customers to orchestrate GPU computing resources for AI and deep learning workloads. At the time of Run:ai's announcement to open source, NVIDIA had just acquired the startup.
Both NVIDIA and Run:ai have handled the acquisition relatively quietly. In April of this year, local and overseas media in Israel reported that NVIDIA was set to acquire Run:ai for $700 million. NVIDIA confirmed the acquisition but did not disclose the scale of the deal, stating instead that Run:ai had been closely collaborating with NVIDIA since 2020. In other words, Run:ai, founded in 2018, began its close relationship with NVIDIA two years after its inception.
In April, NVIDIA stated that it would continue to offer Run:ai's products "under the same business model" and invest in Run:ai's product roadmap as part of NVIDIA's DGX Cloud AI platform. This platform allows enterprise customers to access computing infrastructure and software, which they can use to train generative AI and other forms of AI models.
Media reports have pointed out that Run:ai's software, which made virtualization and orchestration of AI workloads possible, is a necessary capability for effectively allocating GPUs. NVIDIA hopes to integrate this technology into the NVIDIA DGX Cloud system to provide enterprise users with more options for managing AI workloads.
On Monday, NVIDIA did not explain why it plans to open source Run:ai's software. A spokesperson for NVIDIA simply stated in a statement, "We are excited to welcome the Run:ai team to NVIDIA."
Some commentators believe that the reasons behind Run:ai's software being open sourced may not be difficult to understand. Since becoming a leader in AI chips, NVIDIA's market value has surpassed $3 trillion this year, overtaking Apple as the highest-valued company. While this is certainly good news for NVIDIA, it has also faced antitrust regulatory hurdles in acquiring Run:ai. In other words, the decision to open source may stem from regulatory pressure.
Wall Street Journal noted that the U.S. Department of Justice and the European Commission had previously initiated investigations into whether NVIDIA's acquisition of Run:ai would harm competition. Just over a week ago, the European Commission announced its approval of NVIDIA's acquisition, determining that even if the acquisition is completed, other software options available in the market are compatible with NVIDIA products, thus the acquisition would not create a monopoly or hinder competition in the AI and GPU fields