Conversational AI entered everyday life rapidly after ChatGPT demonstrated how accessible large language models could be. Robotics, however, is still pursuing a similarly transformative leap--one that makes intelligent machines practical and widely useful beyond specialized settings.
At TechCrunch Disrupt 2026, Nvidia Inception's Global Head of Physical AI, Les Karpas, is set to explore the barriers separating today's robotic systems from broad real-world adoption. The discussion will focus on how physical AI can become more adaptable, reliable and scalable.
The data challenge in physical AI
A key difference between language AI and robotics is the availability of training data. Language models benefited from vast bodies of digital text, while robots must learn from complex, changing physical environments. Every movement, surface, object and interaction adds new variables that machines need to understand.
Autonomous driving systems have built extensive datasets through years of real-world operation. General-purpose robots face a wider challenge: they must navigate diverse spaces and perform a broad range of tasks safely and consistently.
To address this gap, companies are developing simulation environments, synthetic datasets and foundation models that can learn across different robot designs. These methods aim to help machines gain useful experience before they operate in factories, homes, hospitals or cities.
Nvidia's perspective on intelligent machines
Through its work with startups across robotics, manufacturing, mobility and smart-city technologies, Nvidia is closely involved in the expanding physical AI ecosystem. Karpas brings a multidisciplinary background spanning design, manufacturing, entrepreneurship and investment, reflecting the blend of skills required to build capable robotic systems.
The next major robotics advance may depend less on a single machine and more on shared data, scalable training environments and software that connects digital intelligence with physical action. As these building blocks mature, robots could evolve into more flexible partners across work and daily life.
Physical AI's progress could redefine how people and machines collaborate, turning robotics from a specialized technology into a more accessible part of modern life.
