NVIDIA has introduced a self-improvement program for robots that delegates training direction to teams of AI coding agents rather than human engineers. The system enabled robots to learn precise physical tasks, including installing GPUs and cutting zip-ties. The approach signals that agentic AI paradigms developed for software are now being applied to embodied robotics training pipelines.
Allen Institute for AI has released MolmoMotion, a new model that adds language-guided 3D motion forecasting to the open-source Molmo family. By conditioning spatial trajectory predictions on natural language, the system enables more flexible, human-interpretable motion anticipation. The work targets applications in robotics, video understanding, and embodied AI where predicting movement in 3D space is safety-critical or operationally essential.
A Hugging Face blog post co-authored with Amazon demonstrates how to take AI models from the Hugging Face Hub all the way to running on physical robots. The integration combines Amazon's open-source Strands Agents agentic framework with Hugging Face's LeRobot robotics library to create an end-to-end pipeline. The result is a practical path for developers to deploy Hub-trained policies and models onto real robot hardware using agent-based orchestration.
The article title suggests a discussion of bringing BEV, or bird’s-eye-view perception, into embodied intelligence. It appears to frame robot data as a scaling bottleneck and points to a cross-dimensional approach for accelerating data use. Because no body text is provided, the specific method, company claims, benchmarks, and product details cannot be verified.
NVIDIA and LG Group announced an AI factory collaboration spanning robotics, autonomous driving, data center technologies and GPU cloud services. The effort connects NVIDIA Isaac, Cosmos, DRIVE, DSX, Blackwell GPUs, NeMo and TensorRT-LLM with LG’s manufacturing, robotics, mobility and infrastructure businesses. The partnership also supports LG’s EXAONE sovereign AI model work and broader enterprise AI adoption across the group.
Daxiao Robot and CUHK MMLab introduced Kairos-Homeworld, an open project with 300,000 Chinese residential floor plans and 5,000 interactive 3D home scenes. It can generate full household environments from prompts, including layouts, furniture, objects, and physical properties. The article frames it alongside Kairos 3.0-4B as part of a broader embodied AI stack: world model, data, and environment.
NVIDIA and LG Group are collaborating on an AI factory to support LG’s AI-driven businesses across robotics, autonomous driving, data center technologies and GPU cloud services. The effort connects NVIDIA’s AI factory platform with LG’s manufacturing, mobility, robotics and infrastructure capabilities. It also covers Isaac, Cosmos, DRIVE, DSX and EXAONE-related work using Blackwell GPUs, NeMo, Nemotron datasets and TensorRT-LLM.
NVIDIA and Doosan Group are expanding their partnership across physical AI, robotics and AI factory infrastructure. The collaboration connects NVIDIA’s accelerated computing stack, DSX, MGX and physical AI tools with Doosan’s industrial automation, power generation and electronics materials capabilities. Key areas include smarter industrial robots, autonomous equipment, AI data center power systems and advanced PCB materials for high-performance servers and networking.
At Computex 2026, NXP focused on Physical AI and introduced its Neural Axis architecture for edge devices. The architecture emphasizes low latency, high security, and hardware-based trust for real-time responses. The article frames this as important for robotics, autonomous vehicles, and other physical-world AI deployments where safe operation is essential.
Hugging Face Blog announces NVIDIA Cosmos 3, described as the first open omni-model for Physical AI reasoning and action. The title indicates a focus on AI systems that interact with physical-world scenarios rather than only text generation. Because the article body was not provided, its architecture, supported modalities, license, downloadable assets, benchmarks, and deployment requirements cannot be verified from the available material.
Hugging Face published a tutorial for running Reachy Mini conversations without cloud audio processing or API keys. The setup uses its speech-to-speech library as a cascaded VAD, STT, LLM, and TTS pipeline exposed through a Realtime API-compatible WebSocket. Recommended defaults include llama.cpp with Gemma 4, Silero VAD, Parakeet-TDT, and Qwen3-TTS, while allowing swaps to vLLM, MLX, Transformers, or hosted Responses API providers.
Ars Technica reports that Hugging Face has introduced a roughly $2,500 bipedal humanoid robot project built around 3D-printable legs. The effort targets builders and researchers rather than mainstream consumers, lowering the hardware barrier for hands-on robotics experiments. Its broader significance is in open, reproducible embodied AI research, where models and control systems need physical platforms for testing.
Humanoid robot startup Figure AI recently launched a highly buzzworthy technology showcase: a 24-hour uninterrupted live stream depicting its latest humanoid…
In this episode of the Latent Space podcast, the hosts and guest host Noah Smith (author of the well-known economics and technology blog Noahpinion)…
In this issue of Import AI 451, author Jack Clark opens with a thought-provoking question: "Is there any technological genie that has been released that can be…
Google DeepMind has published a new technology called D4RT, designed to enable artificial intelligence to understand and reconstruct the dynamic world we live…
NVIDIA and Hugging Face have jointly announced the launch of the new Cosmos Reason 2 model, marking a major breakthrough in the fields of Physical AI and…
This article from the Hugging Face blog reveals the latest collaborative breakthrough between NVIDIA and Hugging Face in the fields of Embodied AI and physical…
The cloud AI model deployment and hosting platform Replicate has officially announced support for running the new lightweight vision-language model (VLM) —…
As healthcare demands increase and medical staffing shortages worsen, the development of medical robots — such as robots for ward supply delivery, assisted…
As embodied AI develops rapidly, deploying powerful robotics foundation models onto specific hardware has become a key challenge. NVIDIA and Hugging Face have…
In the history of artificial intelligence, the appearance of the ImageNet dataset in 2012 is widely recognized as the key catalyst that ignited the deep…
At NVIDIA GTC 2025, NVIDIA unveiled a remarkable set of new open-source models and datasets for the field of "Physical AI" — also known as embodied…
Physical Intelligence, a Physical AI startup founded by robotics luminary Sergey Levine and others, has officially open-sourced its flagship robot foundation…
### Background and Challenges In robotics (such as Embodied AI), imitation learning and reinforcement learning require collecting large volumes of robot…
Pollen Robotics has announced the launch of an open-source project called "Pollen-Vision," a unified vision interface designed specifically for robotics…