Real-time Analytics
VNPT needed a high-performance traffic-monitoring system that could analyze thousands of camera feeds and hundreds of thousands of AI service requests to improve congestion management and road safety at busy intersections.
Reported impact
VNPT reduced model-training time from weeks to days, improved traffic-flow visibility, and built a safer road-monitoring system that can help authorities respond to incidents more quickly and efficiently.
VNPT AI built a traffic-monitoring platform on NVIDIA AI Enterprise so it could analyze thousands of camera feeds and handle hundreds of thousands of AI service requests across busy Vietnamese intersections. Video streams are ingested through NVIDIA Metropolis and the NVIDIA DeepStream SDK, where pretrained models, TensorRT optimization, and containerized AI Enterprise components detect vehicles and pedestrians, interpret traffic signals, and classify potential hazards. Training runs on NVIDIA DGX A100 systems, while inference is split between NVIDIA Jetson devices at the edge and on-premises servers equipped with NVIDIA T4 and A30 GPUs to balance low-latency roadside processing with centralized multi-stream analytics. The result is a hybrid edge-plus-datacenter architecture that feeds real-time congestion and safety insights to response teams and gives VNPT a reusable base for future AI services.
Traffic-camera feeds are processed with Metropolis and DeepStream, edge systems with Jetson interpret signals and track road users, backend T4 and A30 servers aggregate multiple streams, and the resulting alerts and congestion insights are delivered to authorities and response teams.
Reduced training time from weeks to days
The video analytics applications are deployed at the edge with NVIDIA Jetson or on premises with T4 and A30 GPUs.
VNPT used NVIDIA Metropolis to process and analyze multiple video feeds from strategically placed cameras around some of the most traffic-heavy intersections in major cities.
they needed extremely accurate, high-performance AI models to handle hundreds of thousands of AI service requests on thousands of cameras across Vietnam.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
NVIDIA Sample Aug 25, 2025 Singtel Singtel needed to make its 5G enterprise offering compelling for AI-heavy customers, but real-time video analytics, AR/VR, and chatbot workloads required low-latency edge compute, shared GPU capacity, and… NVIDIA A40 GPUs · NVIDIA AI Enterprise · NVIDIA T4 GPUs NVIDIA Sample Oct 9, 2026 California Department of Transportation (Caltrans) Caltrans needed to inspect and monitor roughly 1,200 active transportation projects more frequently and safely, but manual site visits were slow, inconsistent, and risky for staff working around traffic, bridges, and… Remote operations up to 100 miles Esri ArcGIS · NVIDIA CUDA · NVIDIA Jetson Orin NVIDIA Sample Sep 30, 2026 EPRI EPRI needed to run advanced AI for sensitive utility documents and weather-planning workflows locally, because energy-sector research data could not leave controlled environments and shared compute delays limited how… 150 tokens/sec aggregate across 10 concurrent Power Chat users DGX Spark · Dell Pro Max with GB10 · NVIDIA CBottle 