Nomadic, a startup developing infrastructure for analysing large-scale autonomous vehicle data, has raised $8.4 million in a seed funding round led by TQ Ventures, with participation from Pear VC and Jeff Dean.
The round values the company at $50 million post-money and will support continued development of its platform, which transforms raw sensor and video data into structured, usable datasets for training and monitoring autonomous systems.
What The Company Does
Nomadic builds a platform that converts large volumes of unstructured video data generated by autonomous vehicles and robotics systems into searchable, structured datasets. Using a combination of vision-language models, the system enables companies to analyse footage more efficiently without relying on manual review.
Autonomous systems generate extensive amounts of data, much of which remains underutilised due to the difficulty of processing it at scale. Nomadic addresses this by allowing teams to query and extract relevant scenarios directly from their datasets, including rare edge cases that are critical for improving model performance.
The platform supports use cases such as fleet monitoring, compliance analysis, and the creation of targeted datasets for reinforcement learning. By enabling automated data structuring, it reduces the need for human annotation while accelerating iteration cycles in model development.
Advertisement
Market Context / Industry Background
As development of autonomous vehicles and robotics systems advances, the volume of data generated by these systems has increased significantly. Training and validating physical AI models requires access to diverse and high-quality datasets, particularly those capturing rare or complex scenarios.
Traditional approaches to data labelling and annotation rely heavily on manual processes, which are time-consuming and difficult to scale. This has led to the emergence of AI-driven data infrastructure tools that automate the identification and classification of relevant events within large datasets.
Companies such as Scale, Kognic, and Encord are also developing automated labelling solutions, while larger players are introducing open-source models to support similar workflows. The shift toward model-driven data processing reflects a broader need for more efficient infrastructure in physical AI development, particularly as deployment requirements continue to grow across industries globally today.
Founder / Investor Commentary
Mustafa Bal, co-founder and CEO of Nomadic, highlighted the importance of extracting meaningful insights from existing data, noting that much of the value lies in helping companies better understand their own footage rather than relying on external datasets.
Varun Krishnan, co-founder and CTO, described the platform as an “agentic reasoning system,” explaining that users can define what they are looking for while the system determines how to identify and contextualise those events across large datasets.
Schuster Tanger, partner at TQ Ventures, emphasised the strategic value of specialised infrastructure, suggesting that autonomous vehicle companies benefit from focusing on core product development rather than building internal data processing systems.
Growth Plans / Use Of Funds
The funding will be used to expand Nomadic’s customer base and continue refining its platform capabilities. The company is also developing additional tools to improve understanding of physical interactions in video data, such as modelling vehicle behaviour and interpreting robotic movements.
Future development will focus on extending the platform to support non-visual data sources, including lidar and multi-sensor inputs, with the aim of providing a more comprehensive data analysis framework for autonomous systems.
Nomadic has already secured customers including Zoox, Mitsubishi Electric, Natix Network, and Zendar, indicating early adoption among companies developing autonomous technologies.
About Nomadic
Nomadic is an AI infrastructure company focused on analysing data from autonomous systems. Founded by Mustafa Bal and Varun Krishnan. Headquartered in the United States. The company develops tools that convert large-scale sensor and video data into structured datasets, enabling faster and more efficient training of autonomous models.
