DeepX presented an open-source alliance for a shift to physical AI at CES 2026. DeepX said on Jan. 9 it hosted the CTA-led studio event "CES Foundry" at CES 2026. At the event, DeepX presented an open-source-based physical AI alliance and ecosystem collaboration as a way to overcome the limits of data centre-focused AI.

The event was held under the theme "Unstoppable Rise of Physical AI". Global companies participating in mass production and commercial services discussed practical challenges and solutions for physical AI.

Hyundai Motor Robotics Lab, Baidu, Edge AI Foundation, Wind River and Ultralytics joined the panel. The panelists said lower dependence on data centres, low power use and low heat, and stable 24-hour continuous operation were emerging as essential requirements in industrial settings such as robotics, smart factories and edge IT services.

Hyundai Motor Robotics Lab Managing Director Hyun Dong-jin said, "Robotics is now core infrastructure that moves society and industry." He said, "The foundation that makes this possible is physical AI, and on-device AI that makes its own judgments is essential even at sites with unstable networks." He added, "Through cooperation with DeepX, we have secured on-device AI for robots that has been verified in real operating environments, and from 2026 we plan to apply it in earnest to next-generation robots and security solutions."

The panelists pointed to differing toolchains by hardware, inconsistent performance metrics and repeated re-optimisation during model porting as hurdles to commercialisation.

DeepX proposed an open-source-based physical AI alliance to address the problem. Partners stressed the need for ecosystem collaboration, saying the success of physical AI depends on how quickly and easily a plug-and-play environment can be created.

DeepX Chief Executive Kim Nok-won stressed that physical AI is becoming a reality. Kim said, "The centre of gravity of AI is shifting from data centres to the physical world," and added, "Hardware and software integration is needed to accelerate this trend."