Analyst(s): Olivier Blanchard
Publication Date: July 7, 2026
NVIDIA announced Halos for Robotics, a full-stack safety system for robotics and physical AI. Agility is the first company to incorporate elements of the platform into its Digit humanoid robot, extending safety capabilities across factories, warehouses, and logistics operations.
What Is Covered in This Article:
- NVIDIA announced Halos for Robotics as a full-stack safety system for robotics and physical AI.
- The platform combines NVIDIA IGX Thor, Holoscan Sensor Bridge, Halos OS, Halos Core, the Outside-In Safety Blueprint, and the Halos AI Systems Inspection Lab.
- Agility is integrating NVIDIA IGX Thor and Halos Core into its proprietary safe human detection system for Digit.
- NVIDIA Halos AI Systems Inspection Lab supports inspection and certification preparation against standards including IEC 61508, ISO 13849, and ISO/IEC TR 5469.
- The Halos ecosystem brings together software vendors, embedded system providers, semiconductor companies, industrial application developers, and certification bodies.
The News: NVIDIA announced Halos for Robotics, a full-stack safety system for robotics and physical AI that unifies AI compute, safety software, sensor connectivity, safety applications, inspection, and certification preparation. The platform extends NVIDIA’s autonomous vehicle safety foundation into robotics, humanoids, autonomous mobile robots, and industrial environments where robots increasingly operate alongside people.
Agility is the first company to incorporate elements of NVIDIA Halos for Robotics into its humanoid robot Digit, integrating NVIDIA IGX Thor and Halos Core into its proprietary safe human detection system. NVIDIA also made Halos Core for NVIDIA IGX available in early access and released the open-source Halos Outside-In Safety Blueprint in early access on GitHub.
How NVIDIA is Building a Critical Safety Layer for Physical AI
Analyst Take: NVIDIA Halos for Robotics gives physical AI a safety architecture at a time when robots are moving from structured industrial workflows into more dynamic environments alongside people, equipment, and other robots. NVIDIA is not treating robot safety as a feature within a single machine, but as a full-stack system that spans AI compute, sensor connectivity, safety software, external perception, inspection, and certification preparation. According to Futurum’s 1H 2026 Intelligent Devices Market Sizing & Five-Year Forecast, Robotics carries the highest percentage CAGR of any destination at 9.9%, making safety, validation, and certification increasingly important as autonomous systems move into real-world industrial environments. The platform connects IGX Thor, Holoscan Sensor Bridge, Halos OS, Halos Core, the Outside-In Safety Blueprint, and the Halos AI Systems Inspection Lab into a common architecture for developing, validating, and deploying physical AI systems. The larger question is whether robotics developers and industrial operators will converge around a common safety architecture as physical AI deployments expand across real-world environments.
Safety Moves Beyond the Robot Itself
Traditional industrial automation typically operated within defined environments where safety controls could be built around predictable workflows and physical separation. The environments targeted by physical AI are different because humanoids, autonomous mobile robots, and other intelligent systems increasingly operate alongside workers, equipment, and moving assets. NVIDIA Halos for Robotics addresses this challenge through a layered architecture that combines industrial-grade AI compute, sensor connectivity, safety operating functions, AI-based safety applications, and inspection processes. The platform builds on more than 18,600 engineering years of autonomous vehicle safety development, along with more than 21 billion safety transistors assessed, over 7 million lines of safety-assessed code, more than 22,000 platform safety monitors, and more than 330 autonomous vehicle safety research papers. The result is an effort to extend safety beyond individual devices and establish a framework that spans the full lifecycle of physical AI systems.
Certification and Inspection Become Part of Deployment
One of the more significant aspects of the announcement is the emphasis on inspection and certification readiness as part of the development process rather than a final step before deployment. NVIDIA Halos AI Systems Inspection Lab is an ANAB-accredited ISO/IEC 17020 Inspection Body focused on AI and functional safety for autonomous vehicles and robotics. The lab enables partners to assess the integration of safety, AI safety, and cybersecurity requirements across systems built on the Halos platform before seeking certification from organizations including TÜV Rheinland, TÜV SÜD, UL Solutions, SGS, exida, and CertX. NVIDIA positions this process as a way to help developers focus certification efforts on their own application logic instead of reevaluating the underlying platform. As physical AI systems move into industrial environments, the ability to streamline inspection and certification processes could become as important as advances in compute or AI models.
Outside-In Safety Expands Environmental Awareness
The Outside-In Safety Blueprint reflects NVIDIA’s view that onboard sensors alone may not always provide sufficient visibility for complex industrial environments. The blueprint uses external infrastructure cameras, AI perception, safety monitoring, event integration, and decision-making systems to extend robot awareness beyond the machine itself. NVIDIA demonstrated the approach through an automated trailer loading use case in which external cameras track people, forklifts, and activity around loading zones to inform safety decisions. The architecture includes the Sensor Input Processing Pipeline, Safety AI Monitor, Safety Event Integrator, and Safety Decision Maker operating on IGX’s dedicated Functional Safety Island. By combining external infrastructure intelligence with onboard safety systems, NVIDIA is attempting to address situations where traditional inside-out safety approaches may limit efficiency or struggle to interpret complex environmental conditions.
Agility Provides an Early Validation Point
Agility’s participation gives NVIDIA an early example of how Halos for Robotics can be implemented in production-oriented environments. The company is integrating NVIDIA IGX Thor and Halos Core into its proprietary safe human detection system for Digit, a humanoid robot designed for logistics, manufacturing, and warehouse operations. Agility will also participate in the NVIDIA Halos AI Systems Inspection Lab to evaluate Digit’s safety-related software, AI components, and cybersecurity protections against standards including IEC 61508, ISO 13849, and ISO/IEC TR 5469. The deployment context includes customers such as Amazon, GXO, Schaeffler, and Toyota Motor Manufacturing Canada, placing the discussion within active industrial workflows rather than controlled demonstrations. While a single deployment does not establish an industry standard, it provides an early indication of how developers may use a common safety architecture as physical AI systems move into larger-scale operations.
What to Watch:
- Adoption rates for NVIDIA Halos for Robotics among robotics developers building systems for industrial environments.
- The extent to which inspection and certification readiness become requirements rather than differentiators in physical AI deployments.
- Early deployment results from Agility’s integration of IGX Thor and Halos Core into Digit.
- Operational benefits delivered by the Outside-In Safety Blueprint in warehouse and logistics workflows.
- Industry alignment around standards such as IEC 61508, ISO 13849, and ISO/IEC TR 5469.
- Participation growth across the Halos ecosystem of software providers, semiconductor vendors, robotics companies, and certification organizations.
- The balance between infrastructure-based safety architectures and traditional onboard sensing approaches as autonomous systems scale.
See the complete press release on NVIDIA Halos for Robotics on the NVIDIA website.
Declaration of generative AI and AI-assisted technologies in the writing process: This content has been generated with the support of artificial intelligence technologies. Due to the fast pace of content creation and the continuous evolution of data and information, The Futurum Group and its analysts strive to ensure the accuracy and factual integrity of the information presented. However, the opinions and interpretations expressed in this content reflect those of the individual author/analyst. The Futurum Group makes no guarantees regarding the completeness, accuracy, or reliability of any information contained herein. Readers are encouraged to verify facts independently and consult relevant sources for further clarification.
Disclosure: Futurum is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of Futurum as a whole.
Read the full Futurum Group Disclosure.
Other Insights From Futurum:
NVIDIA Q1 FY2027: Data Center Diversification, Blackwell Scale, CPU Upside
Will NVIDIA Investment Accelerate Marvell’s XPU Growth?
Author Information
Olivier Blanchard is Research Director, Intelligent Devices. He covers edge semiconductors and intelligent AI-capable devices for Futurum. In addition to having co-authored several books about digital transformation and AI with Futurum Group CEO Daniel Newman, Blanchard brings considerable experience demystifying new and emerging technologies, advising clients on how best to future-proof their organizations, and helping maximize the positive impacts of technology disruption while mitigating their potentially negative effects. Follow his extended analysis on X and LinkedIn.
