Nomadic Raises $8.4 Million to Build the Intelligence Layer for Physical AI Systems

Nomadic, a San Francisco startup, announced it has closed an $8.4 million seed round to develop a specialized intelligence layer for physical AI. Led by TQ Ventures, the round included participation from Pear VC, BAG Ventures, and Predictive VC, alongside tech luminaries like Google DeepMind’s Jeff Dean and Scott Wu. The company plans to use the capital to scale its platform, which helps robotics and autonomous vehicle teams make sense of the vast amounts of video data they collect.

Founded by Mustafa Bal and Varun Krishnan, the team brings experience from industry giants like Microsoft, Snowflake, and Lyft. They observed that while robotics fleets generate petabytes of data, engineers often struggle to find specific, safety-critical events within those archives. Nomadic addresses this by treating video as a medium for reasoning rather than just a source for static labels.

The platform utilizes advanced Vision-Language Models (VLMs) and agentic AI to allow users to describe complex scenarios in plain language. Instead of manual tagging, the system automatically retrieves matching events and validates them using deep motion tracking and segmentation. This hydraulic mining approach ensures that detected behaviors are defensible and ready for use in training or monitoring workflows.

Early adoption is already taking place across several high-stakes sectors, including construction and autonomous driving. Companies such as Zoox and Mitsubishi Electric are utilizing the technology to surface edge cases that would otherwise remain buried. These partnerships demonstrate the platform’s ability to accelerate development cycles for large-scale physical systems.

Unlike traditional labeling services that focus on high-volume annotation, Nomadic positions itself as a reasoning abstraction layer. This focuses on the long tail of failures where physical AI systems are most vulnerable to ambiguity. By providing a way to continuously surface what matters, the company aims to bridge the gap between raw sensor data and reliable autonomy.

Looking ahead, the $50 million post-money valuation provides a significant runway for hiring world-class AI talent. Nomadic is currently recruiting research engineers and designers to build out its distributed GPU inference pipelines. As the industry moves toward more complex real-world deployments, the company is set to become an essential infrastructure provider for the physical AI revolution.