Deep Blue AI compares the approaches of Fei-Fei Li's World Labs and Yann LeCun's AMI Labs to world models: the former builds explicit 3D scenes that are geometrically consistent and can be imported into physics engines, while the latter uses JEPA to make predictions in an abstract representation space.
Within 72 hours, Tesla, OpenAI, and Google made consecutive moves: Tesla's Q3 earnings call stated that Optimus V3 will enter mass production by the end of 2026 and be delivered externally in 2027, and a production line with an annual capacity of 1 million units is already running at the Fremont factory; OpenAI's official website posted 19 robotics hardware positions, covering actuator electromagnetic design, gear design, motor control, and precision transmission. The author argues that embodied intelligence is shifting from an independent track to a natural spillover of general large-model capabilities, and that embodied AI startups that rely solely on financing and lack real-world scenario data feedback will face repricing.
In his GTC Berlin speech, Jensen Huang said the ChatGPT moment for robotics is approaching rapidly rather than having already arrived, and on the same day released upgrades to the Isaac platform, a humanoid robot reference framework, and simulation and data tools to prepare the infrastructure for that moment.
NVIDIA CEO Jensen Huang said at the Build a Claw event in London that OpenClaw is one of the most important models to date, calling it "the first truly trustworthy agent runtime framework" and describing it as providing a real-time runtime framework for LLMs that can be used in robotics. He also stressed that autonomous systems must treat every object as untrusted by default and grant only minimum privileges, and introduced OpenShell, an isolation system NVIDIA developed for this purpose that continuously monitors for privilege violations and issues alerts.
Entire industries will be built on open models. Developers need the freedom to build them. @JensenHuang on why it is important to keep these models open and accessible.
In a 100-minute podcast, Jensen Huang revealed two new growth paths Nvidia is building: charging enterprises software licensing fees of several thousand dollars per GPU per year, and a revenue-sharing model with cloud service providers that rely on its full technology stack. Both depend on the platform's installed base rather than hardware sales.
Nvidia is expanding its Halos operating system from autonomous driving to robotics, allowing developers to customize safety features without disrupting the underlying architecture. The system continuously monitors hardware and software libraries; Holoscan Sensor Bridge can be embedded in microcontrollers or FPGAs to identify corrupted sensor data. It also includes virtual environment simulation and an inspection lab program, providing safety feedback for robot development.
Ken Goldberg of UC Berkeley argues in an article for the Agentic Robotics (AR) approach, in which multi-agent systems write, test, and iterate structured robot control programs offline, as a third path between model engineering and data-driven methods that does not rely on large-scale robot demonstration data.
At its Advancing AI event, AMD introduced the Ryzen AI Embedded X100 series processors for humanoid robots, the Kria AI SOM modules, and the Robotics Software Suite, with sampling in June and mass production in Q4. The article's analysis holds that AMD's combination of x86 real-time control, unified memory, and FPGA targets industrial quasi-humanoid and dual-track supply chain scenarios; in the short term it is unlikely to shake Nvidia Thor's roughly 68% share of the high-end bipedal segment, and its opportunity lies in capturing the installed base of industrial PCs plus discrete GPUs and serving as a second supplier.
New Scientist published a humanoid robot buying guide, comparing five consumer-oriented models including the Unitree R1, 1X NEO, and Figure 03, and raising four questions buyers should consider first.