NVIDIA has a deep and long-standing investment in functional safety. The company has accumulated over 18,000 engineering years on vehicle safety, assessed more than 21 billion safety transistors, and produced over 7 million lines of safety-assessed code. More than 22,000 platform safety monitors have been developed, 330+ research papers on AV safety published, and 30+ certificates and assessment reports issued. This body of work was built for AVs, one of the most demanding safety domains in engineering. The key insight behind Halos OS is that this proven foundation doesn’t need to be rebuilt f
Inside NVIDIA Halos for Robotics: A Full-Stack Functional Safety System for Physical AI | NVIDIA Technical BlogNVIDIA IGX Thor is an enterprise-ready platform for physical AI. It offers server‑class AI performance together with industrial-grade hardware, advanced functional safety capabilities, extended lifecycle support, and an enterprise software stack in configurations suitable for industrial and medical environments. IGX Thor extends this compute and safety foundation to edge systems where uptime, reliability, and standards compliance are central to system design. With the IGX Thor platform, developers can build mission-critical edge computers that operate reliably in harsh physical conditions, int
NVIDIA IGX Thor Powers Industrial, Medical, and Robotics Edge AI Applications | NVIDIA Technical Blog… It incorporates a variety of safety features, including: Hardware fault detection Safe‑state monitoring In‑system test DRAM ECC Freedom from interference FFI Dedicated Functional Safety Island FSI , an independent, redundant safety processor that monitors and handles safety‑critical workloads, prov… …
… This hardware includes: IEC 61508 SIL 3 capable Safety Island FSI : A dedicated functional safety island with up to 12K DMIPs, its own I/O, power, and clocks—physically isolated from the main compute domain High diagnostic coverage : Over 22,000 safety mechanisms provide diagnostic coverage across … …
… She combines her expertise in language modeling and AI evaluation with a commitment to developing trustworthy AI tools that prioritize transparency, safety and responsible deployment. …
… JetPack 7.2 supports two MIG partitions on Jetson Thor: A larger AI and graphics partition for inferencing, rendering, visualization, and general NVIDIA CUDA workloads 12 SMs, 1536 CUDA cores A second isolated compute partition for robotics, control, perception, or safety-critical workloads 8 SMs, … …
… The following sections explain how AI grids enable three such workloads at scale: voice, vision, and media. AI Grid for voice Why Latency is Critical for Voice AI Human-grade voice AI services are extremely sensitive to end-to-end latency. …
… Discuss 0 Discuss 0 Tags Agentic AI / Generative AI | Data Science | Developer Tools & Techniques | General | DGX | Intermediate Technical | News | AI Agent | AI Enterprise | DGX Spark | LLMs About the Authors About Maitri Taneja Maitri Taneja is the product marketing manager for NVIDIA DGX Spark, … …
… In the literature loop, the agent fills a structured worksheet, searches for relevant methods, extracts challenge cards, creates proposal cards, filters duplicates and previously failed ideas, and scores proposals against expected gain, implementation risk, contract safety, evidence, novelty, and r… …
… Unless expressly permitted via a Product Agreement, you agree that your actions and transmission of User Content: a does not include any confidential information; b does not include any controlled or sensitive data, including but not limited to protected health information, personal data, payment c… …