Are Brain Waves the Next Unlock for Physical AI?

What Happened

In a striking development at the intersection of neuroscience and robotics, Encord—a data tooling company specializing in training AI models—has partnered with Zander Labs, a German neuroscience startup, to explore whether brain wave technology can revolutionize how we train physical AI systems. The collaboration took place at Encord's warehouse facility in San Leandro, California, where human operators equipped with advanced brain wave headsets are being used to tag data sets that will train humanoid and warehouse robots.

The trial run involves two key personnel: Andrew Ceja and Sofia Infante, both pilot/robotic trainers at Encord, working under the supervision of Lucas Gehrke, a neuroscientist from Zander Labs. The technology measures brain activity to deduce mental states like error, intent, and surprise—essentially creating a neural interface that can translate human cognitive patterns into machine-readable data.

Vineeth Velmurugan, Encord's head of robot learning and former OpenAI robot lab and Berkshire Grey veteran, oversees the project with ambitious goals. The team aims to build an initial brain wave-tagged data set before scaling operations, but the scale required is staggering: Velmurugan estimates it will take a dataset "something like five times the size of YouTube's video corpus" to break through in robotics training.

Why It Matters

This partnership represents a significant pivot point in AI development. For decades, physical AI—robotics and embodied intelligence—has been considered one of the hardest problems in machine learning. Unlike text or image recognition, where vast datasets have enabled rapid progress, robotics has struggled with the "sim-to-real" gap: models trained in simulation often fail when deployed on physical hardware.

Brain wave technology offers a potential solution by capturing human intent and error patterns directly from neural activity. Instead of relying solely on visual feedback loops or trial-and-error learning, robots could learn from the brain waves of humans performing tasks, creating a more intuitive training method that bridges the gap between human cognition and machine action.

The implications extend beyond warehouse automation. If this technology scales successfully, it could transform how we train autonomous vehicles, medical robots, and even prosthetic limbs—any system where understanding human intent is crucial. The approach essentially treats brain waves as a new form of data input, comparable to text or images but with fundamentally richer semantic content.

What to Watch

The Encord-Zander Labs trial is just the beginning. Several critical questions remain:

1. Scalability: Can this technology move from pilot programs to industrial-scale deployment? The five-times-YouTube dataset requirement suggests massive infrastructure investments will be needed.

2. Regulatory Framework: Brain-computer interfaces face unique privacy and safety considerations. How will regulators classify data collected via neural activity?

3. Technical Challenges: Translating brain wave patterns into actionable robot commands requires sophisticated decoding algorithms that must handle the complexity of human thought processes.

4. Industry Adoption: Will other robotics companies follow Encord's lead, or is this a niche solution for specific applications?

5. Timeline to Commercialization: The current trial focuses on building an initial data set. How long until products reach market?

The partnership between a data tooling company and a neuroscience startup signals that the AI industry is recognizing brain-computer interfaces as more than just consumer gadgets—they could be foundational infrastructure for the next generation of physical intelligence.

By the numbers

• Date: 2026-07-26
• Time: 5:19 PM PDT
• Data set size ratio: five times the size of YouTube's video corpus
• Podcast duration: 37 minutes, 12 seconds

Source snapshot

Sources: - https://techcrunch.com/2026/07/26/are-brain-waves-the-next-unlock-for-physical-ai/