Arm Joins Hands With 80 Companies to Build Physical AI Ecosystem, Sets Common Robotics Standards
TECHWORLD ·
✦ AI Summary
Arm announced the launch of "Arm Total Design for Physical AI" on the 8th.
More than 80 companies are taking part in this initiative across the Physical AI technology stack.
Arm aims to reduce system integration complexity and development risk through this effort, while shortening the time from proof of concept to actual product deployment.
Arm has begun moving in earnest to target the Physical AI market. Physical AI refers to the concept of AI being realized in the physical world. On the 8th, Arm announced the launch of "Arm Total Design for Physical AI," unveiling a collaboration framework for Physical AI development.
More than 80 companies are taking part in this initiative across the Physical AI technology stack. Participating companies include AWS, ECARX, Hugging Face, Liquid AI, NXP, PlusAI, PSYONIC, QNX, Qwen, Siemens, and Unitree Robotics. Domestic participants include Samsung Foundry, Telechips, Gaonchips, and Boss Semiconductor.
The initiative spans the software stack, AI models, sensors, computing hardware, virtual platforms, and digital twins. Arm plans to integrate these capabilities into a single collaborative framework. The main point of Arm's announcement was the launch of "Arm Total Design for Physical AI" and the participation of more than 80 companies.
Through this structure, Arm aims to reduce system integration complexity and development risk. It also seeks to shorten the time from proof of concept to actual product deployment.
Physical AI is a technology that goes beyond the digital space to perceive and judge the real environment before carrying out physical actions. Its applications are expanding to autonomous vehicles, humanoids, industrial robots, logistics, manufacturing, agriculture, and resource development.
Arm said the rise of agentic AI has expanded the capabilities of intelligent systems. It also pointed to extending intelligence into the physical world as a key challenge for future AI technology. Arm cited resource extraction, food production, manufacturing, and transportation as example industries for Physical AI, saying these industries generate several trillion dollars in economic activity worldwide. It also projected an annual computing opportunity of about USD 200 billion in the 2030s.
However, Physical AI is technically more complex than general AI services. That is because it requires AI models with inference performance, real-time processing of sensor data, safety assurance, and precise control of actuators.
Arm argues that the technical challenges of Physical AI require a common foundation for collaboration across companies throughout the technology stack, rather than being solved by individual companies alone. With that in mind, Arm plans to expand the cloud AI collaboration model of the existing Arm Total Design into Physical AI.
Arm identifies perception, AI, real-time control, functional safety, power efficiency, and computing performance as common challenges for autonomous vehicles and robotics. It also cites reducing system integration risk and rapidly developing systems optimized for specific workloads as challenges for OEMs and technology partners. The goal of Arm Total Design for Physical AI is to connect companies across different fields to address these needs.
As an automotive example, Arm developed an integrated digital cockpit reference solution with AWS, Google, HERE, RemotiveLabs, and Siemens. Through this, developers can develop, test, and validate automotive software based on Arm Zena CSS before the actual semiconductor launch.
Arm plans to expand its existing collaboration model across Physical AI. The goal is to provide a development environment where systems can be validated and optimized before actual hardware is built.
Along with expanding the Physical AI ecosystem, Arm also unveiled the Robotics Capability Framework. Arm said robotics technology is advancing rapidly, but there is a lack of consistency in standards defining the level of tasks robots can perform, and there is also a lack of consistency in standards defining the required computing performance, latency, memory, power, and safety.
Richard Grisenthwaite, Arm's chief architect, emphasized the need for a common framework to explain and compare robotic system capabilities in the new manifesto. He cited SAE levels in the automotive industry as a comparison point.
SAE levels are used as a common standard to distinguish the levels of autonomous driving technology. Arm says the robotics sector also needs a common language to describe technology levels and system requirements.
The Robotics Capability Framework is being developed to link real-world use cases with robot behavior and outputs, and to define required system requirements step by step. The system requirements include latency, compute placement, memory constraints, power constraints, deterministic operation, and safety.
Arm reflected a range of views from the robotics ecosystem during the early stages of building the framework. Arm plans to further refine the framework in the future by working with Anaxi Labs, Anyrobotics, FMC³ Robotics, Fourier, GALBOT, Gravis Robotics, Lenovo, McKinsey, and Robotech.ai.
Arm's move is being read as a response to the Physical AI era, an expansion of its IP-centered business model, and a reinforcement of its platform strategy for connecting the entire ecosystem. A Physical AI system cannot be built with an AI model alone; it is composed of a flow that links a sensor's perception of the surrounding environment, the computing system's data analysis, AI's judgment, and the actual operation of real-time control systems and actuators.
Arm is pursuing a strategy to ease technical complexity through ecosystem collaboration in order to lower the difficulty of Physical AI adoption at the ecosystem level rather than leaving it to individual companies. It is also supporting differentiation based on each participating company's areas of strength. Domestic participants include Samsung Foundry, Telechips, Gaonchips, and Boss Semiconductor. Their participation is drawing attention to whether Arm-based Physical AI platforms will expand across South Korea's semiconductor, automotive, and robotics industries.
Arm said that large-scale deployment of Physical AI in the real world requires not only innovative technologies but also a common foundation that the ecosystem can use together. The company said this would accelerate the shift from technological innovation to actual large-scale deployment and help expand new AI-driven opportunities for economic growth.
Source: TECHWORLD · Park Gyu-chan
Original: https://www.epnc.co.kr/news/articleView.html?idxno=406663
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis 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.