AI

Deepx: “CES 2027 Will Mark the Full-Blown Competition in Physical AI Technologies”

TECHWORLD ·

Kim Nok-won, CEO of DEEPX, and Kinsey Fabrizio, CEO of the Consumer Technology Association (CTA), pose for a commemorative photo at the “CES 2027 Media Event” held in San Francisco, U.S., on the 21st (local time). [Photo: DEEPX]

✦ AI Summary

Deepx said CES 2027 will be a stage where physical AI technologies and products emerge in earnest.

CEO Kim Noc-won mentioned the limits of data center expansion amid rising AI demand, along with power and component supply constraints.

Deepx plans to unveil commercialization results for low-power AI semiconductors and customer application cases at CES 2027.

Deepx said CES 2027 will be a stage where physical AI technologies and products emerge in earnest. CES 2027 will be held in January next year.

Deepx said on the 23rd that it had disclosed discussions on changes in the AI industry and major technology trends for CES 2027. Kim Noc-won, CEO and CTO, attended the "CTA Media Event" held in San Francisco, the United States, on the 21st local time.

The event was held as an institutional event hosted by or related to the Consumer Technology Association (CTA). The event was attended by CTA CEO Kinsey Fabrizio, Senior Director of Innovation and Trends Brian Comiskey, global media outlets including Bloomberg, and representatives from major technology companies such as Waymo and Zoox.

The discussion focused on safety and privacy issues arising from the spread of AI, as well as the balance between regulation and industrial development. It also placed major emphasis on the trend of AI expanding beyond the digital realm into the physical world, with the rise of physical AI emerging as a key topic in the process.

Kim said discussions on finding a balance among safety, privacy, regulation, and industrial development on-site are active as AI technology advances accelerate. He also said the pace of AI development among major U.S. AI companies was a point of interest for attendees.

Kim said that as AI demand rises rapidly, data center expansion to respond to it is facing a major challenge in the form of physical constraints. Constraints on data center expansion included the need to secure power, limitations in transformer supply, and supply constraints for key equipment and components such as printed circuit boards (PCB).

Regarding this situation, there was an assessment that simply continuing to expand data centers has limits in meeting rising AI demand. Kim said the importance of technologies that implement AI computing in the physical world with less power and fewer resources will grow, and that the ability to implement more intelligence in the physical world with less power and fewer resources is becoming more important, with high-efficiency physical AI at the center of that change.

The physical AI architecture proposed by Deepx is a method of dividing AI computing between data centers and on-site devices. Under this structure, data centers handle AI training, complex inference, and overall system orchestration, while on-site devices such as robots and machinery perform the necessary inference directly.

Kim presented the potential of this physical AI and said it aims for 100 times the efficiency of data centers. He also predicted that CES will become the central stage where physical AI technologies and products appear in force, and that the same trend will continue in the CES 2027 keynote.

On the day, CTA announced Waymo co-CEO Tekedra Mawakana as a keynote speaker for CES 2027. Waymo is currently providing more than 500,000 fully autonomous driving rides per week, and at CES 2027 it plans to introduce its strategy for pursuing safety and business expansion at the same time.

As Stellantis participates in the CES 2027 keynote, Deepx plans to unveil products from global customers and industrial application cases at CES 2027. The company will demonstrate commercialization results for low-power AI semiconductors at the event, and the focus of the exhibition will be on customer product cases equipped with AI semiconductors.

Deepx plans to present cases that solved on-site power, heat, and cost constraints, including the U.S. sixfab "Alphon X5" equipped with DX-M1. "Alphon X5" won the Best of Innovation Award in the Enterprise Technology category at CES 2026. Through CES 2027, Deepx plans to expand the industrial applicability of physical AI based on low-power AI semiconductors and strengthen contact points with global customers.

Source: TECHWORLD · Park Kyu-chan
Original: https://www.epnc.co.kr/news/articleView.html?idxno=407331

References

This article was produced with the help of an automated content generation algorithm.


Source: TECHWORLD

View original

This article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.