Ultralytics Platform and YOLO26

Ultralytics Inc. is a Delaware corporation with offices in Frederick, Maryland, New York, London, Madrid and Shenzhen. It develops the open-source Ultralytics YOLO family of real-time computer vision models, used by millions of developers, and sells commercial licenses and a hosted platform to companies that want to build vision AI without managing infrastructure.

Ultralytics Platform, launched in March 2026, combines dataset upload and import, SAM 3-powered smart annotation for detection, segmentation, pose, oriented boxes and classification, training of YOLO26 and YOLO11 on cloud GPUs or local hardware, and deployment to auto-scaling endpoints in 43 regions with monitoring. Models export to formats such as ONNX, TensorRT, CoreML, TFLite and OpenVINO for edge devices. An Enterprise License removes AGPL-3.0 obligations and allows on-premise, edge or private-cloud deployment. Logos shown include Siemens, Shell, Philips, Renault, Tata Steel and DHL.

• Smart annotation: SAM 3-assisted masks and boxes for detection, segmentation, pose, OBB and classification.
• Cloud training: train YOLO26 and YOLO11 on cloud GPUs or stream metrics from local hardware.
• Experiment tracking: loss, precision, recall, mAP, confusion matrices and run comparison.
• Managed deployment: auto-scaling inference endpoints in 43 regions with monitoring.
• Edge export: ONNX, TensorRT, CoreML, TFLite, OpenVINO and other formats.
• Enterprise License: commercial use without AGPL-3.0 obligations, with on-premise options and SLAs.

Vietnamese system integrators can train custom YOLO detectors for traffic monitoring or license plate detection and deploy them on edge boxes under a commercial license. Electronics and garment factories can build defect detection models from their own images without an in-house MLOps stack. Agritech firms can train drone-image models for crop counting and disease detection.

One workspace replaces separate labeling, experiment tracking and serving tools, so teams move from dataset to deployed model without provisioning infrastructure. A free plan with signup credits lowers the cost of first projects, and cloud training is billed per GPU hour. Smaller YOLO26 models run fast on edge devices; one customer story reports a 28% cut in inference cost with YOLO26.

YOLO models have become a default choice for real-time object detection on cameras, drones and embedded devices, and teams now want integrated tooling to label data, train and ship models quickly. Vietnamese AI startups, system integrators and factories building inspection, traffic and retail analytics increasingly use YOLO, which makes licensing and managed training relevant locally.

Ultralytics Inc. is a Delaware corporation with offices at 5001 Judicial Way, Frederick, Maryland, plus New York, London, Madrid and Shenzhen. Its open-source YOLO repositories report more than 138,000 GitHub stars and 342 million downloads. The platform states ISO/IEC 27001:2022, SOC 2 Type 1 and GDPR compliance. YOLO26 was released in January 2026.