According to the Intelligence Finance app, the research department of the world's largest auto manufacturer, Toyota Motor (TM.US), and the humanoid robot research company Boston Dynamics under Hyundai Motor Company are joining forces to accelerate the development of humanoid robots based on AI technology. This move signifies that the global auto industry leader Toyota is following in the footsteps of Tesla and Chinese new energy vehicle companies, venturing into the forefront of technology with humanoid robots.

On Wednesday, they stated in a declaration that this partnership would combine Toyota Research Institute's expertise in machine-based large-scale behavior model learning with Boston Dynamics' humanoid robot Atlas. They mentioned that the team from Boston's TRI and Boston Dynamics would conduct research on practical use cases of AI training humanoid robots in areas such as human-machine interaction.

Various 'AI humanoid robots' created based on OpenAI's GPT and other AI large models are also known in the tech industry as 'embodied AI' - referring to the comprehensive integration of artificial intelligence with physical entities in the real world, allowing AI to interact not only in virtual environments but also through some physical form (such as humanoid robots, sensors, etc). Embodied AI focuses on perceiving and understanding the world through physical devices and engaging in deep interactions with the environment and humans, fully mimicking human behavior in the real world.

Toyota stated that the company has made significant breakthroughs in AI teaching and robot learning, while Boston Dynamics was acquired by Hyundai as early as 2020 with its robot guard dogs and humanoid robot arms for warehouse replenishment achieving significant commercial success. The major collaboration they announced poses a potential challenge to other intelligent robot participants, such as Tesla's Optimus robot displayed last week, which is Toyota's strongest competitor in the electric vehicle field. Additionally, many humanoid robot startups including Figure AI have attracted significant investor attention and billions of dollars in investment.

It is understood that the collaboration between Toyota and Boston Dynamics will focus on basic research, with an eye on eventual commercial applications, but executives of both companies refuse to disclose specific commercial timelines or project budgets.

"This technology holds great promise for the future of the company," said Toyota's Chief Scientist Gill Pratt in an interview. "The work we are doing in generative AI can be seen as a huge complement to what Boston Dynamics is doing."

Pratt said the ultimate goal is to bring AI-based large humanoid robots to factory assembly lines and homes, especially for elderly care.

Auto manufacturers have been at the forefront of automating more assembly line processes to reduce labor costs and enhance worker safety systems. According to the latest data from the International Federation of Robotics based in Frankfurt, globally, the automotive industry ranks second to electronics manufacturers in deploying robot assistants in factories, with about a quarter of new robot installations in the automotive manufacturing industry in 2022.

Aaron Sanders, Chief Technology Officer of Boston Dynamics, stated that the agility of Boston Dynamics robots combined with Toyota's humanoid behavior technology is a key differentiator in the company's manufacturing of intelligent humanoid robots compared to other companies.

"This will provide a significant stage for many tech companies that are now committing to scaling up humanoid robots worldwide." He said.

Sanders stated that Boston Dynamics plans to deploy humanoid robots focused on specific tasks in modern automotive factories "in the coming years", while collaborating with Toyota to research the long-term goal of multi-task humanoid robots based on artificial intelligence.

The emergence of heavyweight generative AI like ChatGPT is expected to enable humanoid robots to more quickly acquire new skills based on learned behaviors. Just as virtual assistants like ChatGPT evolved from large GPT language models, large behavioral models will also be used for humanoid robot programming, and humanoid robots empowered by models like GPT will experience an exponential improvement in their perception of the world and their level of interaction with it.

"Toyota Research Institute is experimenting with large-scale robot behavioral models to enable humanoid robots to perform multiple tasks." Laszlo Tenzerik, professor at MIT and head of TRI Robotics Research, said the ultimate goal is to program a robot designed to crack eggs so that, with the support of large language model-powered large behavioral models, it can handle any number of other tasks, such as folding shirts.

"We're making good progress in this area," Tenzerik said. "We are moving very quickly and actively to demonstrate this new capability."

"AI Godfather" Huang Renxun: The next wave of artificial intelligence will be embodied AI.

With the emergence of OpenAI's Wen Sheng video AI model Sora, which can understand and simulate the physical world of motion, the more powerful physical world modeling capability and more comprehensive multimodal standard-based reasoning ability of AI models may promote the prosperity and development of the humanoid robot industry.

NVIDIA, the founder and CEO known as the 'AI Godfather', Huang Renxun, recently stated: "The next wave of artificial intelligence will focus on the field of robotics, with one of the most exciting developments being embodied artificial intelligence." "We are advancing the entire NVIDIA robot technology stack, opening up the technology platform for global humanoid robot developers and companies to access, using the most suitable platform, accelerated libraries, and large artificial intelligence models."

In recent years, Nvidia has focused its R&D work and R&D spending on the field of humanoid robots, firmly believing that humanoid robots will be the core application scenario of AI technology. Nvidia announced at the end of July that it will provide a set of services, models, and robot computing platforms for the world's leading humanoid robot developers, AI model developers, and software manufacturers to develop, train, and build the next generation of humanoid robots.

Nvidia's new series of service products include the new NVIDIA NIM microservices and framework for humanoid robot simulation and deep learning, the NVIDIA OSMO orchestration service for running multi-stage humanoid robot massive workloads, and remote operation workflows that support new AI technology and humanoid robot simulation, allowing developers to use a small amount of human real demonstration data to train humanoid robots.

It is understood that NVIDIA's newly launched NIM microservices provide pre-built containers supported by powerful inference software from NVIDIA, enabling humanoid robot developers to shorten the deployment time of robot simulation solutions from weeks to just minutes. Two newly launched NVIDIA AI microservices allow robot development experts to enhance the simulation workflow of NVIDIA Isaac Sim with generative physical AI. NVIDIA Isaac Sim is a robotic simulation engineering reference application built on the NVIDIA Omniverse platform.

A recent forecast report released by Markets And Markets shows that the global humanoid robot market is expected to reach only about $2 billion by 2024, and is expected to rapidly expand to $13.25 billion by 2028, with a compound annual growth rate of over 45%. Musk made a startling statement at the Tesla shareholder meeting in June, saying that the 'Optimus Prime' humanoid robot could become a core catalyst for Tesla's market cap to rise to $25 trillion.

In August of this year, many leading players in the technology industry such as OpenAI, Microsoft, NVIDIA, as well as tech giants like Amazon founder Bezos, supported the AI humanoid robot startup Figure AI, which launched the second-generation humanoid robot Figure 02. Just as its name implies, this arguably most advanced humanoid robot currently is the successor to Figure 01 released in 2023.

Figure AI's new humanoid robot fully integrates the multimodal AI big model launched by OpenAI. The technology industry generally regards this humanoid robot as a "walking ChatGPT," with many powerful abilities such as mimicking human behavior, deep learning and thinking, and natural and efficient communication with humans.

According to Figure AI, this is the most powerful and comprehensive humanoid robot the company has ever released and combines the flexibility of the human body with the most advanced artificial intelligence model from OpenAI. This humanoid robot's powerful abilities in mimicry and deep learning mean that it can accurately perform various complex and dangerous tasks in enterprise production and manufacturing activities, and play a more intelligent role in assisting humans to improve production efficiency. In the near future, it is expected to fully penetrate households and become a consumer electronics product with a popularity rate comparable to that of iPhones and iPads.