Nvidia announces first full-stack safety system for physical AI
22 June 2026 
Nvidia has announced what it is calling “industry’s first full-stack a safety system for robots and physical AI”. The Halos for Robotics system unifies AI computing and safety. The American humanoid developer Agility is the first company to use the system to build safety into its Digit humanoids that are working for customers including Amazon, Schaeffler and Toyota. The system aims to bring a new standard of responsible automation to factories, warehouses and logistics operations.
Nvidia argues that traditional safety technologies built for structured environments do not function as spaces become more unstructured, and robots move out of cages. The new system brings together powerful AI computing in a single platform with safety.
The safety stack will allow users to rely on a standardised, unified safety architecture that connects AI computing, system software, sensor data, safety applications and inspection for robotic systems. It is aimed at next-generation autonomous robots that will operate in dynamic environments alongside humans, using AI foundation models, accelerated computing and distributed sensors. Nvidia says it will give a single common safety architecture to machines that sense, decide and act in the real world.
“Physical AI is transforming how factories, warehouses and logistics operations work, and robotics teams need a unified safety architecture to scale autonomous systems into these environments,” explains Deepu Talla, Nvidia’s vice-president of robotics and edge AI. “With Nvidia Halos for Robotics, developers and system-builders can harness Nvidia’s proven autonomous vehicle safety foundation to develop safer robots faster, and bring them into industrial operations alongside workers with greater confidence.”
The new system draws on more than 18,600 engineering years of autonomous vehicle safety development. It provides developers with a common safety architecture for building, validating and deploying physical AI systems.
Humanoid robots have to operate in dynamic environments alongside workers, equipment and other robots that are constantly in motion. That requires safety for every layer of the stack. The new system covers the key layers needed for robot safety:
- IGX Thor and Holoscan sensor bridge provide industrial-grade AI computing, built-in safety, and sensor connectivity for real-time robotics and safety workloads.
- Halos OS provides the software stack for robotics safety, including support for safety-related operating functions and safety applications and extending robot perception using external cameras and AI agents to control robot behaviour dynamically.
- The Halos AI Systems Inspection Lab – the first program to be accredited by the Ansi National Accreditation Board (Anab) for functional and AI safety for physical AI. It is helping to prepare Halos integrations for third-party certification by bodies including TÜV Rheinland, UL Solutions, TÜV SÜD, exida, SGS and CertX. More than 40 companies including manufacturers, certification bodies and safety vendors are working to move safe physical AI systems from design to real-world deployment.
“As AI-enabled robotics moves into industrial environments, the industry needs standardised, internationally recognised frameworks to assess safety across increasingly complex systems,” says Ansi’s president and CEO, Laurie Locascio. “Anab’s accreditation of the Nvidia Halos AI Systems Inspection Lab confirms the program has the competence and impartiality to evaluate robotic AI systems against recognised safety requirements, giving companies a rigorous and internationally recognised foundation for their path to certification.”
Agility is teaming up with Nvidia to integrate IGX Thor and Halos Core into its human detection system for its Digit humanoid robot. IGX Thor will deliver industrial-grade AI computing with built-in safety, while Halos Core will support the software layer for safety-related operating functions.
Agility will also participate in Nvidia’s Halos AI Systems Inspection Lab to ensure that Digit’s safety-related software, AI components and cybersecurity protections comply with standards such as IEC 61508, ISO 13849 and ISO/IEC TR 5469 before final third-party certification.
“For humanoids to deliver value at scale, safety has to be built into the robot and validated across the entire system,” says Agility CEO, Peggy Johnson. “Partnering with Nvidia to implement and optimise the Halos for Robotics system, extends our leadership in responsible automation, which is a non-negotiable requirement for bringing humanoids safely into industrial workflows. This collaboration unlocks true human-robot teamwork, driving the long-term returns that will power next-generation manufacturing and logistics operations.”
The Nvidia Halos for Robotics ecosystem brings together partners from industry, sensors and silicon, and certification to support safety from development to deployment:
- Industrial applications Fort Robotics, Inventec, Kion Group, Lyte AI and Neurealm are developing functional safety agents using Nvidia’s Halos Outside-In Safety Blueprint.
- Sensors and silicon Infineon, NXP, STMicroelectronics and Texas Instruments are contributing sensor, safety microcontroller and other semiconductor technologies.
- Software Acontis, FreeRTOS and QNX are supporting the real-time operating environment, safety communications and embedded software layers needed for functional safety development.
- Embedded systems Advantech and NexCobot are delivering safety-designed IGX-based systems for robotic deployment.
- Assessment agencies TÜV Rheinland is inspecting IGX Thor, Halos OS and Holoscan sensor bridge for functional safety certification readiness, building on TÜV SÜD’s inspection of Thor SoC and Halos Core for ISO 26262
