Samsung Electronics Co., Ltd.

Collaboration targets DX-M2-based robotic AI infrastructure, expanding into VLA and VLM technologies for next-generation intelligent robots

SEOUL, South Korea, April 21, 2026 /PRNewswire/ -- DEEPX (CEO Lokwon Kim), a pioneer in ultra-low-power AI semiconductor technology, and Hyundai Motor Group's Robotics LAB have announced a strategic collaboration to jointly develop a next-generation Physical AI computing platform for advanced robotics.

The two parties will co-develop a next-generation AI computing architecture capable of running large-scale generative AI models in real time within robotic systems, with the goal of building an integrated Physical AI platform purpose-built for robotics.

This partnership goes beyond a simple technology exchange — it represents a strategic collaboration to jointly architect the core computing infrastructure for next-generation robot platforms.

The robotics AI field is increasingly centered on Vision-Language-Action (VLA) and Vision-Language Model (VLM) technologies — enabling robots to perceive their surroundings through cameras, understand natural language commands, and make autonomous decisions.

These technologies are considered essential to the evolution of robots from simple automated machines into intelligent systems that can see, understand, and act.

To enable stable deployment of these next-generation AI capabilities in real-world robotic environments, the two companies will collaborate across the following areas:

• Ultra-low-power AI semiconductor architecture
• AI computing hardware systems for robotics
• Physical AI software stack
• Robotics application AI libraries

The ultimate goal is to build a unified AI computing platform capable of running large-scale AI models in real time within robotic systems.

At the center of this collaboration is DEEPX's next-generation AI chip, the DX-M2.

Designed to run large-scale AI models in ultra-low-power environments, DX-M2 is a Physical GenAI semiconductor being developed as the computing platform for real-time AI inference in physical AI environments — including robotics, autonomous mobile systems, and industrial automation.

Its ultra-low-power, high-performance AI computing architecture enables on-device AI inference within robots, expected to significantly enhance robotic autonomy and response speed.

While AI has historically been driven by data center infrastructure, the industry is now entering the era of Physical AI — where intelligent systems extend into robots, industrial equipment, and autonomous mobile platforms.

In this landscape, robots and humanoids are emerging as the most critical Physical AI platforms, directly interfacing AI with the real world.

The Physical AI semiconductor market is projected to reach approximately $123 billion by 2030, with robotics and humanoids expected to be the primary demand drivers.

DEEPX and Hyundai Motor Group Robotics LAB have been jointly developing 'edge brain' technology for robotics based on low-power AI semiconductors over the past three years. This new collaboration expands that foundation into the domain of next-generation robotic AI.

Industry observers anticipate that the robotics sector will develop a new industrial ecosystem centered on standardized computing platforms — similar to what occurred in the semiconductor industry. The emergence of large-scale manufacturing infrastructure combined with Physical AI computing platforms is expected to reshape the structure of the entire robotics industry.

"The AI industry is rapidly shifting from data center-centric models to a Physical AI era, where intelligence expands into real-world systems. The next five years will be decisive in reshaping the industry. In the Physical AI era, ultra-low-power computing technology capable of running AI in real-world systems — robots, vehicles, industrial devices — will become the core infrastructure. DEEPX aims to become the global leader in Physical AI computing platforms that power AI across robotics and industrial systems."

— Lokwon Kim, CEO, DEEPX

"In the era of Physical AI, robots are becoming the closest point of contact between AI technology and people. At Hyundai Motor Group Robotics LAB, our goal is to create robots that can naturally coexist with humans — robots worth living with and working with. To achieve this, we are strategically building a core technology ecosystem — including on-device AI computing — in collaboration with specialized partners across industries worldwide."

— Dong Jin Hyun, Vice President, Head of Robotics LAB at Hyundai Motor Group

View original content to download multimedia:https://www.prnewswire.com/news-releases/deepx-and-hyundai-motor-group-robotics-lab-partner-to-develop-next-generation-physical-ai-compute-platform-for-robotics-302747850.html

SOURCE DEEPX

Global adoption of Physical AI chips accelerates as industry demand expands across robotics, smart factories, and edge AI — backed by a growing partner ecosystem and an early global supply chain spanning Asia, North America, and Europe.

SEOUL, South Korea, March 27, 2026 /PRNewswire/ -- DEEPX, a Seoul-based fabless semiconductor company developing ultra-low-power AI inference chips for physical AI applications, has secured 27 commercial purchase orders across eight countries within seven months of starting mass production of its first-generation AI chip — a pace that industry observers describe as highly unusual for an emerging fabless company at such an early stage of commercialization.

The orders span seven major Physical AI application domains, including robotics, smart factories, edge AI servers, industrial AI, surveillance, AI IT services, and smart cities, with deployments now active across Asia, North America, and Europe.

Global Expansion at Unusual Speed

The ramp-up was slow at first: only two orders came in during the first five months after production began. But the company added 25 additional orders in less than three months in 2026, indicating rapidly accelerating commercial adoption.

While some semiconductor startups have achieved limited international shipments, DEEPX's simultaneous expansion across multiple countries and application domains at this pace is highly unusual, according to industry observers. The company's chips are now being deployed across eight countries spanning Asia, North America, and Europe.

Early Global Supply Chain Execution

DEEPX established a global distribution network early in its commercialization phase through partnerships with Avnet, DigiKey, and WPG. Industry analysts note that it is rare for an emerging fabless semiconductor company to secure a global supply chain at such an early stage of mass production.

In Europe, DEEPX signed a distribution agreement with Avnet Silica — the European arm of Avnet — establishing a regional supply chain ahead of broader market entry. Avnet Silica has already identified more than 30 prospective customers across high-performance embedded segments including smart city infrastructure, autonomous mobile robots (AMR), machine vision, and smart factories. Both companies are actively expanding purchase contracts as European industrial demand for edge AI inference accelerates.

Partner Ecosystem: From Silicon Vendors to AI Software Leaders

A key driver of DEEPX's commercial momentum is a cross-industry partner ecosystem spanning semiconductor vendors, industrial computing platforms, and open-source AI software leaders — reflecting a deliberate strategy to position its silicon as an embedded AI standard rather than a point solution.

Renesas Electronics: Has integrated DEEPX NPUs into more than three types of industrial boards combining Renesas application processors, delivering solutions applicable to smart factory deployments. The collaboration between a tier-one AP semiconductor company and an AI chip startup has been widely noted as an uncommon signal of hardware ecosystem maturity.

Sixfab and Raspberry Pi: Have jointly developed and released an AI HAT module built around the DX-M1, alongside a real-time smart traffic analysis solution using the CES 2026 Best of Innovation-winning 'ALPON X5' AI PC that integrates DEEPX products.

AAEON (ASUS subsidiary), IEI, WeLink, Endrich, Toradex, and Lanner: Industrial system developers have deployed customized AI hardware solutions for smart cities, automated logistics, and security systems powered by DEEPX chips, extending the company's reach across diverse embedded application segments.

Ultralytics & Network Optix: Ultralytics, developer of the widely-adopted YOLO vision AI model family, has demonstrated 'Open-Source Physical AI Alliance' workloads running on the DEEPX NPU with a one-click deployment flow. Network Optix has validated an intelligent Video Management System capable of managing thousands of camera feeds simultaneously using DEEPX inference.

New Products: DX-M1M and DX-AIPlayer

Alongside its commercial order momentum, DEEPX has launched two mass-production product lines designed to reduce integration friction for industrial developers across its key markets:

DX-M1M (M.2 Module): A standard M.2 form factor module embedding the DX-M1 AI chip, designed to drop into existing industrial PCs and edge servers. The format allows customers to add powerful AI inference capability without redesigning hardware, significantly lowering the barrier to deployment.

DX-AIPlayer: An edge AI acceleration solution integrating the DEEPX NPU with a high-efficiency CPU board. Positioned as a 'one-stop' platform covering the full workflow from model development to industrial deployment, the product enables developers to implement vision AI applications without complex integration overhead.

Large-Scale Global PoC Strategy

The company attributes its rapid adoption to a large-scale pre-production engagement strategy. In the semiconductor industry, it typically takes 9 to 18 months from PoC initiation to full production deployment in customer applications. To shorten this cycle, DEEPX began working with customers well before mass production began.

Over more than one year prior to production, the company collaborated with approximately 350 global companies, supporting proof-of-concept testing and system integration. This approach created a large pipeline of customers already familiar with the silicon and ready to transition into commercial orders once production began.

Physical AI Moving from Concept to Reality

CEO Lokwon Kim pointed to the growing volume of PoC activity as the clearest signal that physical AI is no longer a forward-looking concept. "Many people still view Physical AI as a concept for the future. But when we look at the growing number of PoC projects with global companies today, it becomes clear that Physical AI is already becoming a reality," he said.

On the commercial order milestone, Kim was direct. "Securing 27 commercial orders within seven months is one of the early signs that this transition is already happening in the industry."

With distribution in place across Asia, North America, and Europe, and a partner ecosystem spanning silicon vendors, industrial platform makers, and open-source AI software leaders, DEEPX appears positioned to expand its commercial footprint across multiple regions simultaneously. If the current order velocity holds, the company's ambition to establish its NPU as the default physical AI inference chip for embedded developers worldwide may be closer to reality than the industry expected.

View original content to download multimedia:https://www.prnewswire.com/news-releases/deepx-secures-27-commercial-orders-across-8-countries-within-7-months-of-mass-production-302727200.html

SOURCE DEEPX