{"version":1,"type":"story","url":"https://digestai.news/story/nvidia-open-sources-run-ai-platform-under-apache-2-0-license","json":"https://digestai.news/story/nvidia-open-sources-run-ai-platform-under-apache-2-0-license.json","markdown":"https://digestai.news/story/nvidia-open-sources-run-ai-platform-under-apache-2-0-license.md","slug":"nvidia-open-sources-run-ai-platform-under-apache-2-0-license","headline":"Nvidia open sources Run:AI platform under Apache 2.0 license","summary":"Nvidia announced that the entire Run:ai platform is now open source under the Apache 2.0 license on GitHub, making the GPU scheduling, quotas and workload management stack freely available. The release adds Run:ai to the Nvidia DSX AI Factory platform and lets teams inspect, fork and self‑host the code without a paid subscription.\n\nThe monorepo contains a control plane, cluster services, installer, CLI and MCP server; a local development build compiles about 50 Go services. Run:ai runs on Kubernetes or OpenShift clusters with Nvidia GPUs, and a local dev environment can be started on macOS or Linux with a script that creates a 5‑node Kind cluster (Kubernetes v1.32.5) and simulates eight Tesla‑A10 GPUs, requiring 4 CPUs, 16 GB RAM and 100 GB free disk (macOS needs at least 12 GB RAM). The community edition does not include Nvidia enterprise support.\n\nRelated projects such as Model Streamer, vLLM and SGLang are shipped alongside; Microsoft’s benchmark on eight Nvidia H100 VMs showed model loading 3x to 6x faster with throughput of 80 Gbps versus about 40 Gbps on the default path. Deployment can be done from source or via the NGC Community Product Helm chart, and partners including Crusoe, Mirantis, Nebius, OpenNebula Systems, Rafay, Saturn Cloud, Spectro Cloud and vCluster are listed.","keyPoints":["Nvidia released the full Run:ai platform as open source under the Apache 2.0 license on GitHub.","The monorepo includes a control plane, cluster services, installer, CLI and MCP server; local builds compile about 50 Go services.","Run:ai runs on Kubernetes or OpenShift and can be developed locally with 8 simulated Tesla‑A10 GPUs, requiring 16 GB RAM and 100 GB disk."],"whyItMatters":"Open‑sourcing Run:ai lets organizations manage shared GPU clusters without a paid subscription, accelerating AI workload scheduling and reducing vendor lock‑in.","category":{"slug":"enterprise","name":"Enterprise & Industry","url":"https://digestai.news/category/enterprise"},"entities":{"companies":["NVIDIA","Microsoft","Crusoe","Mirantis","Nebius","OpenNebula Systems"],"models":[],"people":["Asif Razzaq"]},"firstPublishedAt":"2026-10-17T17:00:00Z","updatedAt":"2026-10-17T17:00:00Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"marktechpost.com","title":"NVIDIA Open Sources Run:ai Under Apache 2.0 for Kubernetes GPU Orchestration","url":"https://marktechpost.com/2026/10/18/nvidia-runai-open-source-gpu-orchestration-kubernetes","publishedAt":"2026-10-17T17:00:00Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Nvidia open sources Run:AI platform under Apache 2.0 license\", 17 October 2026, https://digestai.news/story/nvidia-open-sources-run-ai-platform-under-apache-2-0-license","publisher":"Digest AI","title":"Nvidia open sources Run:AI platform under Apache 2.0 license","datePublished":"2026-10-17T17:00:00Z","url":"https://digestai.news/story/nvidia-open-sources-run-ai-platform-under-apache-2-0-license"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}