Key points
- PitchBook data shows robotics and physical AI companies raised approximately $48 billion year-to-date.
- Developers are installing cameras in homes, offices, and factories to gather training data.
- This data collection aims to fill gaps in robotics training material using human task completion.
PitchBook data indicates that robotics and physical AI companies have raised approximately $48 billion year-to-date, according to a report by Rafe Rosner-Uddin in the Financial Times. This significant capital influx coincides with aggressive data collection strategies employed by developers seeking to train their systems. The financial activity highlights the growing investment in physical AI technologies that interact with the real world.
By the numbers
- $48 billion raised year-to-date by robotics and physical AI firms according to PitchBook data.
- Cameras installed in homes, offices, and factories to capture human task data for training.
- Training material gaps addressed by observing people completing tasks in various environments.
What the figures mean
The $48 billion in funding reflects a strategic shift toward physical AI that requires extensive real-world data. Developers are placing cameras in sensitive environments like homes and offices to record humans performing tasks, according to the Financial Times report. This approach attempts to bridge the gap in available training material for robotics. For managed service providers, this trend raises privacy and security concerns regarding camera deployments in client premises.
The reliance on human-generated data means that operational environments are becoming data sources. Factories and offices are no longer just workspaces but training grounds for AI systems. This blurs the line between operational technology and data collection infrastructure. Security teams must evaluate how this data is stored, processed, and protected. The scale of investment suggests this practice will expand rapidly across industries.
What should you do?
Managed service providers should review client contracts to understand data ownership and usage rights. Organizations must assess the security posture of any camera systems used for AI training. Data retention policies need to align with privacy regulations and client expectations. Security teams should monitor for unauthorized data exfiltration from these recording devices.
What to do and how to stay safe: robotics
- Audit all camera systems in client environments to verify they are configured for operational purposes only and not unauthorized data collection for AI training.
- Review data processing agreements with vendors to ensure clear definitions of data ownership, retention periods, and usage restrictions for any captured video or sensor data.
- Implement network segmentation to isolate camera feeds and associated data storage from critical operational technology networks, reducing the risk of lateral movement in case of a breach.
- Establish strict access controls for any systems that aggregate or process human task data, ensuring that only authorized personnel can view or export the recorded information.
Step-by-step guide: Shadow AI Explained: 8 Questions Answered for Secure Operations
General security guidance from the BitShield News newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
How much have robotics firms raised year-to-date?
Robotics and physical AI companies have raised approximately $48 billion year-to-date, according to PitchBook data cited in the Financial Times report.
Where are developers collecting training data?
Developers are gathering training data by placing cameras in homes, offices, and factories to record people completing tasks, according to the report.
Why are they collecting this data?
The data collection aims to fill a gap in the training material for robotics and physical AI systems, as stated in the Financial Times article.
