Nvidia has released the full software platform of Israeli-founded Run:ai as open source, giving developers, cloud providers and technology companies tools to use graphics processing units (GPUs) more efficiently for AI workloads. The announcement comes two years after Nvidia acquired Run:ai for an estimated $800 million; the deal followed an eight-month regulatory review in the United States and Europe.

Run:ai, founded in 2018, makes software that manages computing resources for AI workloads, including model training and production use. Its platform can allocate GPU capacity dynamically and share it among workloads that do not need an entire chip.

Organizations can examine and adapt the code, use the complete system or adopt individual components. Nvidia is also distributing the software through its developer infrastructure, while cloud and infrastructure providers are integrating it into their offerings.

The release fulfills a commitment made when Nvidia acquired the company. Founders Omri Geller and Dr. Ronen Dar said then that open-sourcing the platform would help developers build AI systems faster and extend its reach beyond Nvidia GPUs.

The release includes the full platform and four standalone components: KAI Scheduler, GPU Fractioning Operator, KAI Resource Management and Workload Map. They cover workload placement, GPU capacity division and allocation among teams. The community product requires neither a paid Run subscription nor Nvidia support entitlement, but it does not include enterprise support. The software remains subject to its open-source licenses and other terms. Nvidia will continue selling commercial versions through its enterprise AI products and supported cloud offerings.

Nvidia has released the full software platform of Israeli-founded Run:ai as open source, giving developers, cloud providers and technology companies tools to use graphics processing units (GPUs) more efficiently for AI workloads. The announcement comes two years after Nvidia acquired Run:ai for an estimated $800 million; the deal followed an eight-month regulatory review in the United States and Europe.

Run:ai, founded in 2018, makes software that manages computing resources for AI workloads, including model training and production use. Its platform can allocate GPU capacity dynamically and share it among workloads that do not need an entire chip.

Organizations can examine and adapt the code, use the complete system or adopt individual components. Nvidia is also distributing the software through its developer infrastructure, while cloud and infrastructure providers are integrating it into their offerings.

The release fulfills a commitment made when Nvidia acquired the company. Founders Omri Geller and Dr. Ronen Dar said then that open-sourcing the platform would help developers build AI systems faster and extend its reach beyond Nvidia GPUs.

The release includes the full platform and four standalone components: KAI Scheduler, GPU Fractioning Operator, KAI Resource Management and Workload Map. They cover workload placement, GPU capacity division and allocation among teams. The community product requires neither a paid Run subscription nor Nvidia support entitlement, but it does not include enterprise support. The software remains subject to its open-source licenses and other terms. Nvidia will continue selling commercial versions through its enterprise AI products and supported cloud offerings.