Universal Robots has introduced Gen 7, a redesigned collaborative robotics platform that combines new robot hardware, sensing, compute and software to support AI-powered automation and physical AI applications in manufacturing.
Industrial robotics is entering a new phase in which the differentiator is no longer simply how accurately a robot can repeat a programmed movement. Increasingly, manufacturers want machines that can perceive their surroundings, process sensor data, adapt to changing conditions and work alongside AI systems.
Universal Robots is positioning its new Gen 7 robotics platform around that transition.
Introduced at the IMTS tradeshow, Gen 7 represents a broad hardware and software redesign of the company’s collaborative robot platform. The system combines new g-Series robot arms with an AI-ready tool interface, a new controller, updated teach pendant and the PolyScope X robot operating system.
The objective is to make AI and advanced sensing easier to integrate into industrial automation without turning cobot deployment into a specialist engineering project.
That matters because physical AI is moving from research and demonstrations toward industrial applications. IDC’s latest research ranks physical AI as the second-highest AI investment priority among organizations for the next 24 months, behind generative AI assistants and copilots. IDC also describes physical AI as a systems market spanning devices, software, infrastructure and services rather than simply another category of robot.
Universal Robots’ Gen 7 reflects that systems-level approach.
Putting sensing closer to the robot
The new g-Series arms include an AI-ready tool flange designed to provide data, power and safety capabilities directly at the robot’s end effector.
That architecture is aimed at reducing the cabling and integration work normally associated with connecting cameras and high-bandwidth sensors to robotic systems. It also incorporates force-torque sensing, impedance control and Real-Time Data Exchange capabilities.
For physical AI applications, these capabilities are important because intelligence depends on more than visual perception.
A robot performing dexterous manipulation or force-sensitive assembly needs to understand physical contact as well as what its camera can see. Force feedback can help determine whether a component has been inserted correctly, whether an object is being gripped properly or whether a task needs to be adjusted.
This brings industrial cobots closer to the perception-action loop associated with physical AI: sense, interpret, decide and act.
Gen 7 introduces three new robot arms—the UR10g-1750, UR17g-1300 and UR18g-950—alongside Universal Robots’ existing e-Series and UR Series products.
The naming convention reflects payload and reach, while the broader platform is intended to provide a common foundation for AI-enabled applications.
More compute at the edge
Gen 7 also introduces the CB7 Core controller, which Universal Robots says delivers 40% more computing power while occupying 30% less space than the previous generation.
The controller supports multi-network communication with systems including PLCs, HMIs, manufacturing execution systems, industrial PCs and vision platforms.
That connectivity is becoming increasingly important as AI moves closer to production equipment.
IDC’s manufacturing research points toward a factory architecture in which generative, predictive and agentic AI increasingly interact with connected assets and operational data. The firm predicts that by 2027, 40% of operational data will be integrated across applications and platforms autonomously as standardization and AI agents increase.
A robot controller therefore becomes more than a motion-control component. It can serve as an edge computing and connectivity point between physical machinery and the software systems responsible for perception, planning and optimization.
PolyScope X becomes the software layer
The software side of Gen 7 is centered on PolyScope X, which provides open APIs and ROS2 communication while supporting external AI processing through industrial PCs.
That openness is strategically important.
Rather than forcing manufacturers into a closed AI stack, Universal Robots is allowing third-party developers and integrators to connect their own computer vision, AI and automation applications.
The platform already supports third-party technologies from companies including Cognex, Schunk, Magswitch, Robotiq and Inbolt.
The Inbolt integration provides an example of where this model can lead. According to the company, its 3D vision guidance can run natively on the UR controller, reducing the need for a separate vision PC and additional cabling.
This is the broader significance of Gen 7: AI capabilities can increasingly become part of the robot platform rather than an external system bolted onto it.
AI needs industrial safety
Physical AI also introduces a requirement that does not exist in the same form for conventional software: the AI system ultimately controls machinery operating around people.
Gen 7 includes collaborative safety capabilities with a PLd, Category 3 architecture under ISO 13849-1. Universal Robots says its g-Series robots are certified to ISO 10218-1 and UL 1740, while Teradyne Robotics has achieved IEC 62443-4-1 ML3 certification for cybersecurity development practices.
Those controls matter because industrial AI cannot be evaluated purely on model accuracy.
Manufacturers need predictable behavior, access controls, cybersecurity, traceability and safety mechanisms. A vision model that misidentifies an object is one problem; an AI-driven robot making the wrong physical movement around an operator is another.
The ecosystem becomes the competitive layer
Universal Robots is showcasing Gen 7 alongside more than 20 ecosystem partners at IMTS. The company’s UR+ ecosystem provides a curated portfolio of automation products that are validated by both partners and Universal Robots.
That ecosystem strategy could become increasingly important as physical AI develops.
IDC’s research suggests the industrial robotics market is evolving from individual hardware products toward broader systems combining robots, AI models, sensing, software and services.
For Universal Robots, an open ecosystem provides a way to benefit from that innovation without developing every vision, gripping, sensing or AI capability internally.
It also gives manufacturers a more modular path to automation.
Instead of replacing an entire production cell, a company can add a cobot, integrate a vision system, connect an AI model and expand capabilities incrementally.
That approach could prove particularly valuable for manufacturers operating high-mix, low-volume environments where conventional fixed automation can be difficult to justify.
The larger industry shift is clear: physical AI is becoming less about putting a general-purpose model inside a robot and more about building the complete systems needed for machines to safely interact with the physical world.
Gen 7 is Universal Robots’ attempt to make that transition practical at the cobot layer.
Market Landscape
The physical AI market is expanding from experimental robotics into industrial applications, with collaborative robots, machine vision, force sensing, edge AI and robotics software converging into integrated systems.
IDC estimates the global robotics market will exceed $40 billion by 2029 and identifies physical AI as a systems-level opportunity spanning platforms, infrastructure, services and devices.
Industrial robots remain an important installed base for physical AI. IDC reports that China’s 2025 industrial embodied-intelligence market was approximately RMB 5.74 billion, with industrial-robot-based applications accounting for roughly RMB 3.62 billion.
Competition includes established automation companies, industrial robot manufacturers, machine-vision vendors, AI-chip providers and emerging physical AI companies. The competitive advantage is increasingly shifting toward the quality of the complete hardware-software ecosystem rather than robot mechanics alone.
Top Insights
- Gen 7 combines new cobot hardware, sensing, compute and software into a platform designed for AI-enabled industrial automation.
- Its AI-ready tool flange moves sensing, data and power closer to the robot’s end effector, reducing integration complexity.
- The CB7 controller provides 40% more compute in a 30% smaller footprint, according to Universal Robots.
- PolyScope X provides APIs and ROS2 connectivity for integrating external AI processing and third-party automation applications.
- Safety, cybersecurity and open ecosystem support are becoming core requirements as physical AI moves into production environments.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
