Los Angeles, CA, August 15, 2025 –(PR.com)– VidRovr today announced it is building the next generation of large‑scale video intelligence to extend their capabilities into smart‑city and retail‑analytics deployments with the NVIDIA Blueprint for Video Search and Summarization (VSS) and Cosmos Reason, a new reasoning vision‑language model (VLM).
Why is now the right time for Video AI at this Scale?
Video already accounts for 80% of global internet traffic and is growing by approximately 20% annually. A single minute of video shot at 30 frames per second conveys the semantic equivalent of 1.8 million words—far more than any human can process at scale. VidRovr’s patented Video AI architecture enables a single reviewer to query thousands of cameras at once, reasoning across time, location, and multiple feeds to deliver actionable insights at superhuman speed and scale.
What VidRovr Will Demonstrate?
Fine-Tuning and Embedding Lookup with Cosmos Reason 1, enabling users to ask complex, natural-language questions across a network of cameras—such as “Show me when a large crowd forms and quickly disperses on the sidewalk” or “Identify vehicle accidents and dangerous driving scenarios.” This accelerates safety, security, and forensic investigations, while also supporting advanced retail analytics—for example, “Show me how long a customer stays in my store if they purchased a big-ticket item like a TV.”
Using NVIDIA Metropolis for high-accuracy video AI, Cosmos Reason 1 achieves over 90% recall on cross-camera events with a compact 7-billion-parameter model that runs seamlessly from edge GPUs to cloud clusters. This enables city-scale and chain-wide deployments without incurring exponential compute costs.
VSS Blueprint enables fast retrieval across weeks of footage, with key-event highlights automatically generated for dashboards, reducing manual review time from hours to minutes.
“Working with NVIDIA supercharges our system with city-wide and site-wide observability. By integrating the NVIDIA Blueprint for Video Search and Summarization from Metropolis and Cosmos Reason 1 into VidRovr’s existing machine learning pipelines, users can now interact with video data in a more natural, freeform way—asking complex questions and getting fast, actionable answers. This unlocks intuitive, real-time intelligence at massive scale.” – Bret Paulson, Chief Executive Officer, VidRovr
About VidRovr
VidRovr is pioneering Generation 3 computer vision—enabling real-time video observability across vast camera networks. With over 25 billion video frames processed and more than 15 machine learning models orchestrated per deployment, VidRovr delivers actionable intelligence in under a minute, at a scale and speed unmatched in the industry. Learn more at www.vidrovr.com.
All product names, logos, and brands are property of their respective owners. NVIDIA and the NVIDIA logo are trademarks or registered trademarks of NVIDIA Corporation.
