Edge Computer Vision - AWS Panorama (Discontinued) and Its Replacement Stack
AWS Panorama ended on May 31, 2026. Learn the replacement stack for on-site AI analysis of existing IP cameras (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 depended on connectivity to the 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. This article remains available as a historical reference and migration guide.
Why Edge Computer Vision Matters
Sending IP camera footage to the cloud for ML analysis presents challenges in terms of bandwidth cost, latency, and privacy. A single HD camera requires 2-5 Mbps of bandwidth, and streaming multiple cameras to the cloud results in enormous network costs. For use cases that demand real-time processing (safety monitoring, quality inspection), the round-trip latency to the cloud is unacceptable. Additionally, video data contains personal privacy information, making cloud transmission itself a risk in some cases. AWS Panorama was a service that deployed ML models to an edge device (the Panorama Appliance) and analyzed IP camera feeds locally in real time. Video data was processed at the edge, and only analysis results (metadata) were sent to the cloud.
Panorama Appliance and Model Deployment
The Panorama Appliance was an edge device equipped with an NVIDIA GPU that could simultaneously analyze up to 8 IP camera streams. It connected to IP cameras via the RTSP protocol, captured video frames, and ran ML model inference. Custom models built with SageMaker or pre-trained models (object detection, person detection, etc.) could be used. Models were packaged as Docker containers and deployed to the edge device through the Panorama console. Application logic (Python) processed inference results and executed actions based on conditions (sending alerts, recording metrics). Model updates could also be performed remotely from the cloud, eliminating the need for on-site visits.
Use Cases and Pricing
Key use cases for Panorama include manufacturing line quality inspection (at the time, by deploying Lookout for Vision models to the edge for real-time defect detection; that service was also discontinued in October 2025), retail customer analytics (counting visitors, analyzing foot traffic, detecting shelf inventory levels), construction site safety monitoring (checking for helmet and safety vest compliance, detecting intrusions into restricted areas), and parking lot occupancy detection (calculating available spaces in real time from camera feeds). The Panorama Appliance hardware cost approximately $4,000 as a one-time purchase, with a monthly service fee billed per camera stream processed (see the pricing section below). Communication costs for sending metadata to the cloud were minimal, and there were no video data transfer costs. Third-party Panorama-compatible devices such as the Lenovo ThinkEdge SE70 were also available, allowing hardware to be matched to the use case.
Panorama Pricing and Operations
With the shutdown, the following pricing no longer exists (kept as a historical record). The Panorama Appliance was a one-time purchase of approximately $4,000. The monthly fee was billed not per device but per active camera stream processed on the device, at $8.33 per stream per month, plus $0.10 per GB per month for cloud storage of deployed application assets (ML models and business logic). Since inference ran at the edge, there were no bandwidth costs for streaming video to the cloud, making the total cost advantageous compared to cloud inference for real-time use cases. Model updates could be delivered OTA (Over-the-Air), allowing inference logic to be improved remotely without visiting the site. Device health and inference performance were monitored with CloudWatch metrics.
Summary - Migration Guidelines
AWS Panorama offered the simplest path to edge AI analysis of existing IP cameras, but it was discontinued on May 31, 2026, and the appliances no longer work. To build an equivalent setup today, run models built with SageMaker on NVIDIA Jetson-based devices or edge servers, distributed and managed with IoT Greengrass. The design of inferring at the edge without sending video off-site remains valid wherever privacy and bandwidth requirements matter.
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.