A startup focused on improving how autonomous systems learn from real-world data has raised new funding to expand its platform, which converts large volumes of video into structured datasets for training and analysis.
Nomadic AI announced an $8.4 million seed round at a $50 million post-money valuation, led by TQ Ventures, with participation from Pear VC and Jeff Dean.
Addressing Data Bottlenecks In Autonomous Systems
Companies building self-driving cars, robotics systems, and autonomous machinery generate large amounts of video data used for training and evaluation. Much of this data remains underutilized due to the difficulty of organizing and analyzing it.
Nomadic AI’s platform uses vision-language models to transform raw footage into structured, searchable datasets. This enables developers to identify specific scenarios, including rare edge cases, and use them to train models more effectively.
Mustafa Bal said many organizations have the majority of their fleet data stored without being actively used, limiting progress in developing autonomous systems.
Platform Capabilities And Use Cases
The system allows users to query video data for specific events, such as unusual driving conditions or rare operational scenarios. These insights can be used for compliance, monitoring, and reinforcement learning workflows.
Varun Krishnan described the platform as an agent-based system that can interpret user queries and identify relevant data by combining multiple models.
Customers including Zoox, Mitsubishi Electric, Natix Network, and Zendar are using the platform to support development of autonomous technologies.
Competitive Landscape And Industry Trends
The approach aligns with a broader shift toward automated data labeling and analysis in physical AI. Companies such as Scale AI, Kognic, and Encord are developing similar tools.
Nvidia has also released open-source models, including Alpamayo, to support these workflows.
Funding Use And Product Development
The new funding will support customer onboarding and continued development of the platform. Nomadic AI is working on additional tools that analyze physical interactions, such as vehicle lane changes and robotic movements captured in video.
The company is also exploring ways to integrate non-visual data, including lidar sensor inputs, to enhance model training.
Technical Challenges And Future Work
Bal said processing large-scale datasets alongside complex AI models presents ongoing challenges, particularly when extracting actionable insights from terabytes of video data.
The company aims to extend its system to handle multiple data types and improve how autonomous systems interpret real-world environments.
Featured image credits: Universal Toyota
For more stories like it, click the +Follow button at the top of this page to follow us.