Live intelligence

Latest AI updates

12 updates
Industry

Genesis AI Raises $500 Million to Deploy Dexterous Robots

Genesis AI has raised $500 million in a Series A round, valuing the startup at $3.5 billion. The company is developing dexterous robots and combines worker-worn data gloves, simulation, and its Eno robot.

Genesis AIroboticsfunding
Models

Nvidia's GR00T 2 learns robot tasks by transferring human hand movements from video

Nvidia has introduced GR00T 2, an AI model designed to teach robots new tasks without laborious teleoperation. The system transfers human hand movements from video to robotic arms and was trained on about 400,000 hours of footage. Success rates vary widely, ranging from 10 to 80 percent depending on the task.

NvidiaGR00T 2robotics
Research

Grasp-Taxonomy Conditioning as a Mode Switch for Dexterous Grasping of Climbing Holds

Researchers introduce BetaGrasp, a 3D diffusion policy that conditions a DP3-style point-cloud policy on a four-class grasp label (crimp, sloper, pinch, and jug) for a Franka arm with a 16-DoF LEAP hand. They also present a climbing-hold benchmark with a spring testbed that measures slip force in newtons. Without the grasp label, simulated rollouts all ended in a jug grasp regardless of hold, achieving 25% correct grasp type versus 94% for BetaGrasp. In 80 paired hardware trials, BetaGrasp raised average slip force from 9.6 N to 14.3 N and reduced complete slip-offs from 44% to 10% of trials.

roboticsdexterous manipulationdiffusion policies
Research

UniOMA: Unified Optimal-Transport Multi-Modal Structural Alignment for Robot Perception

The paper proposes UniOMA, a multimodal alignment method that augments existing contrastive losses with a geometry-aware Gromov-Wasserstein (GW) distance-based regularization. It encodes intra-modal geometry for each modality and reduces cross-modal structural discrepancies through a dynamically learned GW barycenter, encouraging structural consistency beyond instance-level contrastive alignment. The approach is applied to robot perception tasks involving underexplored modalities such as force and tactile signals, where data often exhibit varying sample densities. Experiments across five robotic benchmarks reportedly improve structural alignment and downstream performance over state-of-the-art multimodal contrastive baselines, and the GW regularization is described as integrable into most InfoNCE-like objectives.

roboticsmultimodal representation learningoptimal transport
Research Editor selected

LLM Repairs Robot Policy by Editing Token-Embedding Weights

The paper applies weight-space programming to a robot foundation policy: a language model reads the policy's weights and edits token-embedding rows using a CPU workspace and budgeted simulator feedback, with no gradient updates or new demonstrations. Across three independent sessions on π0.5, success on 765 held-out paraphrases improved, with the best run rising from 86.3% to 92.0% (48 repairs, four regressions, exact paired p=1.31e-10). Gains concentrate on the 335 instructions containing edited tokens (74.3% to 87.5%), while matched random-direction edits do not help. The authors characterize the result as precise lexical repair rather than general acquisition of missing robot capabilities, and separately report that benchmark metadata lowered FastWAM success from 70.0% to 48.6%.

Why it matters: It offers a concrete, unusually detailed demonstration that an LLM can fix a robot policy's instruction grounding by editing a handful of embedding rows, plus a cautionary evaluation finding about benchmark metadata effects.

roboticsmodel editingLLM agents
Research

RM-VLM uses vision-language models to ground LLM-generated reward machines

Researchers propose RM-VLM, a reinforcement-learning approach that has a large language model generate a reward machine from a natural-language task description. A vision-language model then serves as the labelling function, grounding the machine's propositions in image-based observations during policy learning. A VLM-based value estimator assesses progress toward reward-machine transitions and produces a dense reward signal. The method is evaluated on non-Markovian grid-world navigation, robotic manipulation, and 3D object collection, where it reports higher task success rates than existing VLM-based reward methods.

reward machinesvision-language modelsreinforcement learning
Research Editor selected

ID-WM: Hierarchical World Models with Implicit Dynamics for Long-Horizon Planning

Researchers introduce ID-WM, a hierarchical world model that decouples state-space reachability from local control to improve long-horizon planning. A high-level model samples feasible transitions, while a low-level inverse dynamics model grounds them into actions, reducing cumulative errors. In navigation and manipulation benchmarks, ID-WM achieves a 38.5% improvement over state-of-the-art planners and a 135% gain over single-level MPC in complex unseen configurations.

Why it matters: This paper presents a novel hierarchical approach with significant performance gains, offering practical value for embodied AI and robotics planning.

world modelshierarchical reinforcement learningplanning
Research Editor selected

Microsoft Research Demonstrates Offloaded Inference for Physical AI Robotics

Microsoft Research argues that running physical AI inference exclusively on onboard GPUs limits task performance and battery life, and presents a systematic study of mobile manipulation workloads showing offloading improves success rates and lets robots run larger models. The work reports that some smaller GPUs could not accommodate the stack, and that lighter GPUs slowed mapping and planning by up to 383% and reduced obstacle-detection timeliness by 30% versus an A100. Offloading with a Raspber

Why it matters: It offers concrete benchmark-backed evidence and an open-source toolchain for a robotics infrastructure decision that affects model size, task reliability, and battery life.

Microsoft ResearchPhysical AIinference infrastructure
Industry Editor selected

ZhaoGuan Electronics Relocates HQ to Hangzhou to Advance Physical AI and Embodied Vision

Hangzhou&#39;s industrial sector has welcomed a significant new player. In late April 2026, ZhaoGuan Electronics officially relocated its global headquarters to Linping District, Hangzhou, further bolstering the city&#39;s ambition to become a premie<!--autointro-->

Physical AIroboticsedge computing
Industry

AI Edge Inference Chip Startup SiMa.ai Negotiates Funding at $1.4 Billion Valuation

May 18 — AI edge inference chip startup SiMa.ai is currently negotiating a new round of funding, with a post-money valuation of approximately $1.4 billion. SiMa.ai specializes in AI inference chips designed for the edge, with its products primarily t<!--autointro-->

AI chiproboticsindustrial automation