NeuroFrame raises $45M Series A and launches live video understanding SDK

AI · 4 min read

NeuroFrame closed a $45 million Series A led by Horizon Ventures, with participation from media and surveillance-focused funds, to expand its engineering team and scale inference infrastructure. The company says the new capital will accelerate integrations with edge vendors and secure partnerships with telco operators for low-latency streaming.

Alongside the round, NeuroFrame launched a production-grade SDK that delivers multi-object tracking, pose estimation, and semantic scene understanding at sub-100ms latency on modern edge GPUs. The SDK supports C++, Python, and an upcoming WebAssembly runtime, aiming to make real-time video analytics accessible to robotics, retail analytics, and smart city teams.

Design teams at NeuroFrame emphasized developer ergonomics: the SDK includes a visual annotation pipeline, prebuilt dashboards, and adaptive models that can be fine-tuned with a few labeled clips. The startup also announced a pilot program with three major carriers to test real-time analytics at the network edge, positioning itself as a bridge between video AI and distributed compute.