AWS Panorama Is Discontinued - Migration Guide for Edge Computer Vision
AWS Panorama ended on May 31, 2026, and its appliances stopped working. This article covers the official replacement stack (SageMaker, NVIDIA Jetson hardware, IoT Greengrass) and how Panorama worked.
End of Support and Migration Paths
AWS Panorama reached its end of support on May 31, 2026, and had stopped accepting new customers on May 20, 2025. Because the Panorama Appliance and its applications depended on connectivity to the Panorama cloud service, the devices themselves no longer function after the shutdown. AWS officially points to two alternatives: adopt an off-the-shelf solution from AWS Partners, or build your own with Amazon SageMaker (including the computer vision algorithms in SageMaker JumpStart), run it on off-the-shelf NVIDIA Jetson-based devices or edge servers, and manage the edge fleet with AWS IoT Greengrass, AWS Systems Manager, or Amazon EKS / ECS Anywhere. This page remains available as a historical reference and migration guide.
Overview of Panorama
Panorama was a service that deployed computer vision ML models to existing IP cameras and ran real-time inference at the edge. There was no need to send camera footage to the cloud, solving both privacy and bandwidth challenges. A single appliance (equipped with an NVIDIA GPU, approximately $4,000) could process up to 8 camera streams in parallel at 30 fps, sending inference results to the cloud via IoT Core. Video data stayed entirely within the appliance and never left the network, making it well-suited for environments with strict data locality regulations such as GDPR or HIPAA.
Appliance and Model Deployment
The Panorama Appliance was a dedicated device installed on-premises that connected to IP cameras via the RTSP protocol. Object detection models built with SageMaker (e.g., product defect detection) were deployed from the console to the appliance, applying real-time inference to camera feeds. Applications were written in Python and triggered actions based on inference results (sending alerts, stopping production lines). Only inference result metadata was sent to the cloud and monitored with CloudWatch. Deployment was performed OTA (Over-the-Air) via the console or API, enabling the same model to be rolled out simultaneously to appliances distributed across multiple sites. Model version management integrated with SageMaker Model Registry, allowing pipelines that only delivered production-approved models to the edge.
Model Optimization and Multi-Camera Processing
The Panorama Appliance was equipped with an NVIDIA GPU and processed video from multiple IP cameras simultaneously. Compiling models with SageMaker Neo optimized them for Panorama hardware, improving inference speed. Applications were written in Python, implementing frame preprocessing with OpenCV, model inference, and result post-processing (bounding box rendering, alert determination). A single appliance could process up to 8 camera streams in parallel, simultaneously monitoring multiple angles on factory production lines or in retail stores. Inference results were sent to IoT Core via MQTT and could be visualized as CloudWatch metrics or used to trigger alerts via Lambda.
Use Cases and Deployment Scenarios
Panorama's primary use cases converged on three areas: quality inspection in manufacturing, safety monitoring in logistics warehouses, and customer analytics in retail stores. In manufacturing, cameras on production lines captured product appearances and detected defects (scratches, chips, color irregularities) in real time to automatically stop the line. Compared to traditional dedicated inspection equipment, the ability to reuse existing cameras and update inspection logic through software offered significant flexibility. In logistics warehouses, pose estimation models detected proximity between forklifts and workers to prevent collision accidents. In retail stores, heatmap generation models visualized customer traffic patterns for shelf layout optimization. In all cases, not sending video to the cloud achieved both customer privacy protection and communication cost reduction.
Comparison with Rekognition Video and DeepLens
Similar AWS services to Panorama included Rekognition Video and DeepLens (service ended in January 2024). Rekognition Video analyzes video streams in the cloud, allowing use without edge hardware, but incurs bandwidth costs and latency from sending video to the cloud, making it unsuitable for environments with strict privacy requirements. DeepLens was a camera-integrated device for education and prototyping but did not support production workload performance or multi-camera processing. Panorama was positioned as an industrial-grade service with a production-grade GPU, 8-stream parallel processing, and OTA operational management. For scenarios with few cameras and ample bandwidth, Rekognition Video could start without initial hardware investment; Panorama became advantageous when camera counts increased and bandwidth costs or latency became issues.
Panorama Pricing and Deployment Considerations
With the shutdown, the following pricing no longer exists (kept as a historical record). The Panorama Appliance consisted of a one-time hardware purchase cost of approximately $4,000 plus a monthly service fee, billed not per device but per active camera stream processed on the device, at $8.33 per stream per month. Compared to cloud-based video analysis (Rekognition Video), edge processing was more cost-efficient in environments with many cameras. Since video was not sent to the cloud, bandwidth costs were also reduced. During deployment, the camera's RTSP stream URL, network settings (appliance IP address, DNS), and association with the AWS account were configured. A key consideration was that the appliance required internet connectivity for receiving management commands and downloading models, so it could not be used in fully air-gapped environments. Additionally, due to GPU memory constraints, larger model sizes reduced the number of streams that could run concurrently, so the balance between model size reduction and stream count had to be verified during the evaluation phase.
Summary
Panorama added edge AI to existing IP cameras, but it reached end of support on May 31, 2026, and the cloud-dependent appliances stopped working. To rebuild an equivalent setup, deploy models built with SageMaker onto off-the-shelf NVIDIA Jetson-based devices or edge servers, and handle application distribution and device management with IoT Greengrass. The design principle of running inference at the edge without sending video to the cloud, which optimizes both privacy and bandwidth, remains just as valid in the replacement architecture.
References (Official AWS Resources)
The primary sources for this page are the official AWS website and documentation. Check the official pages below for the latest specifications and pricing.
- AWS Panorama official page
- Amazon SageMaker official page
- AWS IoT Greengrass official page
- AWS Documentation (official)
If this page and the official documentation disagree, treat the official documentation as authoritative.