According to sources familiar with the matter, ChatGPT developer OpenAI, backed by tech giant Microsoft, has launched a new wave of recruitment that suggests the $500 billion AI unicorn is seeking to leverage its expertise in AI large models and generative AI applications to further advance the practical deployment and social penetration of humanoid robots globally, aiming to become a leader in the embodied AI trend.

Sources reveal that OpenAI, led by Sam Altman, has recruited a group of top-tier AI application researchers focused on developing control systems for embodied AI. The AI unicorn is particularly focused on developing core large models or complete robotic systems for humanoid or largely human-like robots. The company is especially concentrated on training robotic AI large models that can power humanoid robots capable of handling extremely complex situations in the physical world.

"Embodied AI" refers to the cutting-edge technology field represented by AI humanoid robots, which is the future technology track most favored by NVIDIA founder and CEO Jensen Huang and Tesla Motors founder and CEO Elon Musk. NVIDIA is building humanoid robot "integrated software and hardware" standards with GR00T + Jetson Thor + Isaac, while Musk boldly claims that Optimus (the humanoid robot) will become the main engine of Tesla Motors' long-term value, stating multiple times that Optimus will eventually contribute about 80% of Tesla Motors' value.

Sources indicate that OpenAI has also posted multiple job listings seeking senior engineers focused on robotics. A recent new hire is Chengshu Li, who joined OpenAI from Stanford University this summer. His research involves developing specialized humanoid robots capable of completing multiple household tasks. Earlier this year, OpenAI first published senior robotics engineer positions focused on hardware systems, including a senior EE sensor engineer, a robotics mechanical design engineer, and a TPM specialist manager.

Altman outlined his ultimate vision for humanoid robots, artificial general intelligence, and human society in a June blog post. OpenAI's leader stated: "If we have to build the first million humanoid robots the traditional way, but then they can operate entire vast industrial chains and supply chains—including mining and refining minerals, driving trucks, running factories, etc.—to manufacture more robots, which in turn can build larger-scale chip manufacturing facilities, data center infrastructure, etc., then the pace of human social progress would obviously be completely different."

**Humanoid Robots - The Technology Field Where Musk and Huang Continue to Double Down**

As developed economies globally enter an era of "severe labor shortages," aging populations and service sector labor shortages are forcing automation to move from production lines to more general "humanoid robot forms." Additionally, with AI technology accelerating comprehensive updates and iterations, embodied AI with powerful capabilities including mimicking human behavior, deep learning and thinking, and natural, efficient communication with humans may become the largest-scale carrier of AI technology for human society.

On September 7, global electric vehicle leader Tesla Motors officially launched a Weibo account named "Tesla AI," posting its first content: "I've been working hard to improve my physique," accompanied by images of the new Optimus humanoid robot. From publicly available images, the Optimus 2.5 version appears more human-like in appearance, with smoother joint designs, especially hand configurations that closely approach human flexibility, smoother metal shells, significantly reduced external cables and actuators, and notably optimized overall structure.

Tesla Motors has reportedly begun producing Optimus 2.5 humanoid robots on pilot production lines at its flagship factory in the United States, with progress exceeding market expectations. As Tesla Motors releases its "Master Plan 4.0," Musk has reiterated the strategic position of the robotics business at Tesla Motors. He stated that scaling FSD and Optimus would be the most important matters, as approximately 80% of Tesla Motors' future value will come from Optimus robots.

Wall Street financial giant Morgan Stanley stated that if Tesla Motors can perfectly master "fully autonomous driving technology" (FSD technology), widespread adoption of humanoid robots is just a matter of time. "Most automotive and robotics experts we've communicated with believe the humanoid robot market is far larger than the current global automotive total addressable market (TAM)," Morgan Stanley's analysis team wrote in a recent research report.

Regarding Optimus production planning, Musk has provided specific targets: Tesla Motors plans to produce thousands of Optimus humanoid robots in 2025, initially conducting production work tests within Tesla Motors factories. "If everything goes well, we plan to increase production 10-fold next year (2026), to approximately 50,000-100,000 units; the year after that, we'll further increase it 10-fold, to 500,000 units and beyond."

In short, if Musk's "new promise" can be successfully implemented, Optimus production in 2027 would be 100 times that of 2025.

Unlike Tesla Motors, NVIDIA, which has continuously increased its investment in humanoid robots in recent years, focuses on being a "computing power and platform supplier for embodied AI"—essentially "tools + operating systems"—coordinating across three lines: cloud training, edge inference, and physical simulation tools. This involves large-scale sales of massive AI GPU clusters needed for robot training/inference systems + edge computing modules + Isaac simulation and toolchains, serving humanoid robot players across industries.

**Huang's Prediction: Following AI, Robotics Will Become NVIDIA's Strongest Growth Engine**

Huang has stated more than once this year that following AI large models, robotics technology, especially humanoid robots, will be the largest potential growth market for this AI chip super giant. "We have many major growth opportunities company-wide, with AI and robotics being the two largest, with total addressable markets reaching trillions of dollars," Huang said when answering shareholder questions at NVIDIA's annual shareholder meeting.

As mentioned above, NVIDIA focuses on coordinating cloud training, edge inference, and simulation tools across three lines. Therefore, for NVIDIA's performance prospects, the humanoid robot era will bring continuously expanding demand for cloud training and inference clusters: humanoid robots require massive multimodal data and joint "perception-control" training, driving sustained upward demand for Blackwell and more advanced architecture-level AI GPU clusters; edge computing shipment expansion will also be a major trend: Jetson AGX Thor targets humanoid body inference (VLA/body balance/grasping/multi-sensor fusion), becoming "GPU computing devices on robots," forming end-to-end NVIDIA ecosystem solutions with GR00T and other models.

NVIDIA is committed to integrating new NVIDIA robotics technology based on 3D comprehensive dynamic simulation and real-world simulation of the physical world into tech companies' robot model development and training processes. Therefore, NVIDIA's Isaac/Isaac Sim/Omniverse platform is expected to become the industrial standard for "simulation + data generation + robot strategy verification," with platform subscriptions and developer ecosystems expected to bring high-stickiness recurring revenue data and increasingly strong moats based on integrated software and hardware.

"We are moving toward an era where the world will soon have billions of robots, hundreds of millions of fully autonomous vehicles, and tens of thousands of super robot factories that can be powered by NVIDIA's software and hardware ecosystem technology," Huang stated.

A recent forecast report published by Markets And Markets shows that the global humanoid robot market size is expected to be only about $2 billion in 2024, potentially expanding rapidly to $15.26 billion by 2029, with a compound annual growth rate expected to exceed 39%. The research institution stated that starting in 2024, humanoid robots will widely penetrate multiple industries including home care, education, healthcare and surgical assistance, automotive manufacturing, and high-risk manufacturing, potentially significantly improving operational efficiency across industries.

International bank Deutsche Bank stated in a research report that humanoid robots will see large-scale robot production and widespread application within the next decade; Deutsche Bank's analysis team predicts that by 2035, the humanoid robot market size is expected to reach $75 billion, and by 2050, the market size could reach $1 trillion, with global sales potentially exceeding 70 million units.

According to Deutsche Bank's analysis report, 2024 will mark the beginning of flourishing application testing of humanoid robots in factories and warehouses, while 2025-2026 will be the critical period for accelerated humanoid robot production. Technology companies focused on humanoid robots, including Tesla Motors, Figure AI, and Agility, have already set aggressive mass production plans for humanoid robots.