AI in Robotics and Physical Systems
Robotics, embodied AI, autonomous systems, manufacturing automation, drones, vehicles, and AI systems acting in the physical world.
A GPT-6 Astra-Assisted Build Turns a Conversational Idea Into Hardware
Justin Lin says GPT-6 Astra helped turn his idea for a voice-controlled Codex pet into a physical device, designing its circuit board from plans he supplied. It was his first project from a bare board to an assembled PCB, and the work involved factory revisions and multiple shell versions. Lin says the process expanded what he felt able to attempt, making him more willing to think bigger.
Lunar Settlements May Need Nuclear Power Where Sunlight Fails
Financial Times science editor Clive Cookson argues that nuclear power could be essential to a permanent lunar settlement because the moon’s most promising resource sites may receive little or no sunlight, and lunar nights can last two Earth weeks. Those same sites would expose reactors to radiation, impacts, extreme heat and abrasive dust, making reliable power a major engineering challenge. Cookson sets the case for nuclear energy against an uncertain commercial prospect: projected revenues for a lunar economy do not establish that a viable market—or a permanent outpost—will follow.
Northwood Raises $375 Million at a $2 Billion Valuation
Northwood has raised $375 million at a $2 billion valuation to expand the ground network that connects satellites, CEO Bridgit Mendler said in an interview with Bloomberg’s Ed Ludlow. Mendler argues that the ground segment remains fragmented across vendors, slowing the deployment of infrastructure as the space industry grows. Northwood is aiming to take on more of that work itself, from building antenna hardware to deploying sites and connecting them with software.
NVIDIA Halos SAFE Connects Road Tests, Simulation, and Hardware for AV Evaluation
NVIDIA argues that autonomous-vehicle safety evaluation should connect measured road tests to repeatable simulation and hardware checks, rather than treat them as separate tests. Its Halos SAFE framework reconstructs captured scenarios with Omniverse NuRec, checks simulated results against real-world measurements and vehicle hardware, then uses generated variations to retest software releases. NVIDIA says the loop can produce evidence for safety targets and certification across markets.
Genie 4 Lets Users Steer Generated Worlds as They Unfold
Google DeepMind presents Genie 4 as a world model built for continued interaction: users can move through generated environments, steer scenes with text prompts and see how they develop in real time. The company’s demonstrations show prompts changing an unfolding scene, from directing a crane toward a piano to adding events and characters to other settings. DeepMind also proposes virtual robot practice as a possible application, while describing access as a limited research preview.
Digital Twin Simulation Aims to Cut Months From Audi Logistics Testing
Audi plans to use a digital twin built in Dassault Systèmes’ DELMIA to test logistics tasks for robots before deploying them in production. The proposed system combines factory knowledge from the 3DEXPERIENCE platform with NVIDIA Omniverse libraries for simulated perception and motion, creating a virtual training environment for robots from Agile Robots and idealworks. NVIDIA says scaling tests to multi-robot fleets could help assess safe operation and save months in future logistics testing.
EKF SLAM Couples Robot Pose and Map Uncertainty
EKF SLAM uses correlations between a robot’s pose and landmark positions so that an observation can refine both the robot’s location and its map. But as landmarks are added, the covariance matrix grows quadratically with the size of the state, making updates more costly. In Stanford professor Marco Pavone’s lecture, unknown landmark correspondences add another risk: a mistaken match can persist in the filter’s later estimates.
AI Factories Are Being Built to Run Agents, Not Just Models
NVIDIA CEO Jensen Huang argues that AI infrastructure is shifting from systems built to run models to AI factories designed to run agents, which he describes as the basis of a new industrial economy. In his GTC Berlin keynote, he extended that case to physical AI, presenting shared platforms and reference systems for vehicles and robots as part of the effort to build Europe’s AI economy.
AI Agents Build Robot Training Worlds From CAD Models and Real-World Motion
NVIDIA presents Agent Astra as a way to automate parts of robot development, turning a prompt, CAD assets and real-world motion data into a simulation for training and testing skills. In the demonstration, Astra builds the world, replays motion and tunes its physics; Isaac Lab and Newton handle training, while Isaac Lab-Arena evaluates skills locally on DGX Station. NVIDIA positions Agile ONE SN, built on Isaac GR00T and powered by Jetson Thor, as a reference robot for bringing trained skills into the physical world.
NVIDIA Links Road Tests, Simulation, and Vehicle Hardware in One Safety Loop
NVIDIA presents Halos SAFE as a framework for evaluating autonomous-vehicle safety by linking road tests, simulation, fleet data and vehicle hardware in a repeatable testing loop. Its central claim is that measured driving scenarios can be reconstructed with Omniverse NuRec, varied across conditions with Cosmos, and reused to test software changes and vehicle responses against safety targets. NVIDIA says the approach is intended to support evaluations from driver assistance through L4 robotaxis and provide evidence for certification across markets.
Xona Raises $250 Million to Build a Commercial GPS Alternative
Xona says GPS’s limitations are becoming more consequential as autonomous vehicles, drones, robotics and other systems rely on precise location with less human oversight. CEO and co-founder Brian Manning argues that smaller, lower-cost satellites in a closer orbit can improve navigation while working with existing GPS devices through software updates. The company has raised $250 million at a $2.45 billion valuation to build and launch its network after receiving FCC approval to operate alongside GPS.
AI-RAN Puts Network Functions and AI Services on Shared Infrastructure
Nokia and NVIDIA argue that telecom operators can turn network infrastructure into a shared platform for radio access and AI workloads, creating a path from programmable 5G toward 6G. Their AI-RAN collaboration pairs Nokia software with NVIDIA accelerated computing; the companies say it can improve network performance, support AI services at the network edge and help automate network operations.
Europe’s AI Strategy Depends on Building Local Compute and AI Factories
In his GTC Berlin keynote, NVIDIA CEO Jensen Huang argued that Europe’s AI economy will depend on producing intelligence, not just adopting AI services. He called for investment across the stack—from energy and computing to AI factories, models and applications—and presented NVIDIA’s hardware, software and partner ecosystem as infrastructure for that production. His case extended from industrial agents and robotics to secure runtimes and local workstations: AI, he said, requires systems built to run models, connect them to knowledge and tools, and put them to work.
Spatial Intelligence Extends AI Beyond Language Into Three-Dimensional Worlds
Fei-Fei Li argues that the next frontier in artificial intelligence is spatial intelligence: systems that can perceive and work with three-dimensional worlds, not just language and flat images. In a conversation with historian Stephen Kotkin, the Stanford computer scientist and World Labs co-founder connects that ambition to a broader question about how intelligence works—and says the development of powerful AI also demands truth-seeking, public responsibility and greater attention to teachers and families.
AI Models and Robotic Labs Narrow Candidates for B-Cell Lymphoma Therapy
Lila Sciences is connecting AI agents to an autonomous laboratory to search for candidate therapies intended to help a patient’s own immune cells fight B-cell lymphoma. The company’s workflow uses Iris, built on NVIDIA Nemotron and Agent Toolkit, to turn a research goal into an experimental plan; models built with NVIDIA BioNeMo then run millions of virtual experiments to narrow candidates before robots make and test them. Lab results feed back into the next round of designs, though the source presents a research process, not a finished therapy.
Tesla’s Flying Roadster Prototype Won’t Reach Production Until 2028
Tesla’s Roadster prototype jumped and flipped using cold-gas thrusters developed with SpaceX engineers, but production remains years away, with a launch not expected until 2028. On Bloomberg Tech, Ed Ludlow described the demonstration and said Elon Musk brought audience members forward to show it was real, not AI-generated; Bloomberg columnist Max Chafkin argued the spectacle was also a pitch for Tesla stock and, potentially, a merger with SpaceX.
NVIDIA’s GTC Berlin Keynote Frames Tokens as a Path From Data to Action
NVIDIA CEO Jensen Huang’s GTC Berlin keynote opening video presents tokens as the building blocks that turn data into knowledge and action. Its examples range from treatment planning and crowd forecasts to robots carrying medical samples and inspecting hazardous sites, connecting AI to decisions and physical work. The video places those examples alongside computing infrastructure and closes with a shared ambition for a future built for all humankind.
AI Agents Link Virtual Experiments to Lab Tests, Factory Plans to Floor Conditions
NVIDIA’s GTC Berlin presentation argues that industrial AI agents need more than a model: they must connect to company knowledge and domain-specific tools, with controls over what they can access and execute. It illustrates that architecture through Lila Sciences’ experiments for therapy design, Siemens’ production planning and Dassault Systèmes’ vehicle-design workflow. Lila feeds physical test results into another round of designs and experiments; in Dassault Systèmes’ example, simulation results prompt refinement before an engineer selects the final design.
Micro Duck’s Walking Policy Trains in the Cloud, Then Returns to Local Simulation
The workflow uses NVIDIA RTX Spark for scene creation and policy evaluation, while cloud compute handles parallel training in Isaac Lab; the trained policy then returns to the local system before deployment to Micro Duck. The operator says Astra uses tools including FreeCAD, Blender and Isaac Sim to build the course and robot model, and that Isaac Lab trains thousands of Micro Ducks in parallel. The account shows the robot walking after deployment but reports no separate measurement of its real-world performance.
NVIDIA Frames AI Compute as Revenue-Producing Infrastructure
NVIDIA’s GTC Berlin 2026 keynote closing video presents AI computing as industrial infrastructure with the potential to generate revenue, linking GPUs and data centers to a wider stack of models, software and robotics. In a musical recap performed by robot characters, the company also calls for agents to act within security limits and urges Europe to build across that stack. The performance makes these points through slogans rather than technical or financial evidence.
KITE Lets Learners Feel Forces in Digital Physics Simulations
Elizabeth Childs, a Stanford PhD candidate, presents KITE as a way to make forces in physics simulations physically perceptible. The hand-moved robot travels freely until a learner encounters a virtual object or force, then pushes back with corresponding force or torque. Childs argues that this combination of tangible movement and responsive feedback lets students explore relationships such as lift, molecular attraction and friction through physical interaction rather than screen representations alone.
Known 3D-to-Image Correspondences Make Camera Calibration Solvable
Camera calibration estimates how a camera maps three-dimensional points to pixels, a mapping that published specifications may not define precisely enough for reconstruction, Stanford professor Marco Pavone argues. Known 3D points and their observed image locations are used to estimate a combined projection matrix, from which the camera’s intrinsic parameters and pose can be recovered. But calibration does not resolve depth from a single image: every point along the ray through a pixel could have produced that projection.
Europe’s Robotaxi Race Runs Through National Rules and Local Training
Europe’s national routes to robotaxi service are advancing at different speeds, but regulatory progress does not make a deployment portable, Frank McCleary, a partner at Arthur D. Little, told Bloomberg’s Ed Ludlow. Operators must train their systems in each new environment, McCleary said, so experience gained in China does not by itself establish that a vehicle will work in a European city. Ludlow framed Europe as an option for Chinese operators facing constrained access to the US; McCleary described the US restrictions behind that constraint.
AI Is Linking Semiconductor Design, Fab Operations and Inspection
NVIDIA argues that accelerated computing and AI can help semiconductor manufacturers manage growing design and fab complexity by linking simulation, digital twins, engineering support, robotics and inspection. Its account centers on a feedback loop: simulation and production data inform models, while inspection results feed back into them. NVIDIA reports chamber simulations up to 35 times faster and more than 96% accuracy in wafer-map defect classification, figures it presents as examples of the approach rather than measures of overall fab performance.