您的浏览器禁用了JavaScript(一种计算机语言,用以实现您与网页的交互),请解除该禁用,或者联系我们。 [凯捷研究院]:物理人工智能 - 发现报告

物理人工智能

报告封面

Taking human-robot collaboration to the next level Table of contents Who should read this report and why? This report is intended for senior executivesshaping their organizations’ approach torobotics and automation. It examines howphysical AI is transforming robotics – from thecapabilities it enables to the value it unlocks,the timelines for adoption, and the barriers thatmust be addressed to scale deployments safelyand effectively. It will be particularly relevant totechnology and innovation leaders (includingchief technology officers, chief innovationofficers, chief digital officers, and heads of AIor robotics), as well as manufacturing, supply chain, and logistics leaders responsible forrobotics strategy and deployment. leaders preparing their organizationsfor wider robotics adoption – includingimplications for governance and risk oversight. As robotics expands into consumer-facing andservice environments – such as healthcare,retail, hospitality, and entertainment – thereport is also relevant to chief product officers,product strategists, and experience designleaders who are shaping interactions betweenpeople and intelligent machines. This report draws on a global survey of1,678 senior executives across 15 industries,complemented by in-depth interviews withindustry experts, robot manufacturers,foundation-model startups, technologyproviders, investors, and academics. Please see the research methodology at theend of the report for more details. In addition, the report provides practicalguidance for CROs and safety or regulatory Key definitionsWe extend our sincere thanks to the many experts fromindustry and academia who shared their insights with us Ashutosh SaxenaFounder and CEO,TorqueAGI Sanjay AggarwalVenture Partner,F-Prime Capital Deepu TallaVP and GM – Robotics &Edge AI,NVIDIA Rebecca Yeung Daniela Rus Strategic Advisor at Dexterity andformer Corporate Vice Presidentfor Operations Science andAdvanced Technology atFedEx Director, Computer Scienceand Artificial IntelligenceLaboratory (CSAIL),MIT Dirk Geiger Milad Malekzadeh Daniel JackerCEO,ZaiNar Angelo Cangelosi Anto Patrex Co-Founder and Vice President AI,Neura Robotics Founder and CEO,CosmicBrain AI Senior Director and TeamLead – Humanoid Robotics,Infineon Technologies Co-Director of the ManchesterCentre for Robotics and AI,University of Manchester Viki Yang Julien PerrinCOO,Niryo Jim Ma Nagesh Puppala Pedro Zheng Overseas Sales Director,UBTECH Robotics Senior Regional Sales Manager,Unitree Robotics Regional Technical Director,Unitree Robotics General Manager, Robotics andPhysical AI, Client ComputingGroup,Intel Corporation Executive summary Physical AI takes AI beyond screens into the realworld – enabling machines to perceive, reason,and act autonomously. This report focuses onits application in robotics, where physical AIrepresents a fundamental shift: from robotsthat follow fixed, pre‑programmed paths torobots that can generalize across tasks, perceiveand navigate complex environments, makecontext-aware decisions, and adapt to real-worldvariation. This enables robots to function infar more diverse and dynamic environments,expanding their reach across nearly every majorindustry and unlocking solutions to problemsearlier automation couldn’t address. "The last decade of AI wasabout information. The comingdecade will be about action." Rebecca Yeung Strategic Advisor at Dexterity and former Corporate Vice Presidentfor Operations Science and Advanced Technology at FedEx Executive summary Executive summary systems could not handle. In parallel, advances insimulation are shortening robot training cycles,while an AI‑robot‑data flywheel is acceleratingimprovement with every real-world deployment.Combined with falling costs of key hardwarecomponents such as sensors, actuators, andelectric motors, and commercial models suchas robotics-as-a-service (RaaS), these shiftsare lowering barriers to adoption. At the sametime, demographic and economic pressures– including aging workforces and persistentlabor shortages – are intensifying demand forrobotic systems capable of taking on roles thatare increasingly hard to staff. Record venturecapital investment into physical AI and roboticsis adding to the momentum behind these shifts. To understand the impact of physical AI onrobotics and the value it can potentially unlock, thisreport draws on a global survey of 1,678 executivesacross 15 industries, complemented by in-depthinterviews with experts across the physical AIand robotics ecosystem (please see the researchmethodology for more details). A game-changer across multiple dimensions Physical AI marks a step change from earlierautomation. By enabling robots to interpretcontext, adapt in real time, and operate inunstructured environments, physical AI promotesthem from passive tools to active collaborators inthe workspace – opening the door to a reimaginedwork environment, in which humans, robots,and AI agents work in tande