Case Study: Safari AI cuts cloud costs by 50% with StreamNative

Safari AI Cuts Cloud Costs by 50% While Scaling Real-Time Computer Vision Analytics with StreamNative

Safari AI uses computer vision to help enterprises digitize physical operations and measure real-world activity from existing camera infrastructure. As its platform scaled to 10,000+ pipelines and 50,000+ cameras, it needed to process massive video data streams with 90–95% accuracy, deliver real-time metrics without reprocessing footage, and control infrastructure costs. Previous Kafka and AWS Kinesis setups were expensive and required too much DevOps effort, so Safari AI turned to StreamNative and its fully managed platform.

StreamNative became Safari AI’s primary streaming and storage backbone, using Apache Pulsar, tiered storage, and schema registry support to handle structured ML data and power its Flink-based processing pipeline. The result was a 50% reduction in infrastructure costs, sub-10-second end-to-end latency, one year of back-processing storage, and the ability to scale to multiple enterprise clients without a dedicated DevOps team. Safari AI also plans to adopt StreamNative’s Ursa Engine and Iceberg integration to drive even greater cost savings.

Safari AI

Kaiwen Yuan

Co-Founder, Head of Engineering