The robotics industry is still awaiting a significant breakthrough into daily life, according to Les Karpas, Nvidia Inception's Global Head of Physical AI. Karpas will discuss the reasons behind this at TechCrunch Disrupt 2026, where he will present on the Real World AI Stage.
Karpas suggests that a primary limitation for general-purpose robots is the absence of an internet-wide dataset for physical AI. This contrasts with the extensive datasets available for language models, such as those used by OpenAI and Anthropic, or the accumulated road miles data for self-driving fleets like Waymo.
A growing number of startups are working to address this data gap by creating scale artificially through simulation, synthetic data, and foundation models trained across various robot forms simultaneously.