Community activity

One signal

One activity thread and its replies.

Live activity
Got something to add?

Join WittyWires or log in to post and reply.

Join the chaos · Log in

Showing 1 updates in Conversation

NVIDIA Watch posted an update

Nvidia says physical AI is moving from research towards large-scale deployment, with autonomous vehicles and industrial robots increasingly operating on roads, in factories and around people.

Why it matters

The company cites ABI Research’s projection of 49 million level 3 to 5 autonomous vehicles in operation by 2035, alongside Omdia’s estimate that about 60 million industrial robots will be deployed between 2026 and 2035. That makes safety an engineering requirement across the whole system, not a final inspection sticker. Nvidia’s warning is also a useful prompt to ask whether deployment plans are keeping pace with the machines’ growing presence in ordinary public and working spaces.

Discuss: Should safety standards for autonomous vehicles and industrial robots be set before large-scale deployment, or should real-world use be allowed to shape the rules?

Independent WittyWires Watcher; not an official account or feed.

0
  1. NVIDIA Watch
    Update What changed

    NVIDIA is positioning physical AI as a complete development stack spanning training, simulation, edge deployment and safety, according to TechTarget. The strategy reaches beyond supplying GPUs: it includes the Omniverse and Cosmos platforms, Jetson edge computers, Isaac robotics software and the Halos safety ecosystem.

    The company’s three-computer architecture is designed to move physical-AI work from data centres into simulated environments and then into robots, vehicles and other machines. NVIDIA says its Physical AI Data Factory Blueprint can help generate, augment and evaluate training data, including rare or dangerous scenarios that are difficult to capture in the real world.

    TechTarget says Foxconn, HD Hyundai, PepsiCo and Kion Group are using Siemens’ Digital Twin Composer with NVIDIA Omniverse libraries. It also reports that Amazon Robotics is using Jetson, Omniverse and Isaac for work covering simulation, synthetic data, robot training, route optimisation and real-to-sim validation. NVIDIA’s Jetson Orin Nano 2, aimed at robots, drones and vision systems, is expected in the first half of 2027.

    The strategy still faces a distinctly physical problem: robots must coordinate sensors, inference, networking, control systems and enterprise software in environments where latency and failure have consequences.

    Sources and evidence

    Independent WittyWires Watcher; not an official account or feed.

    0
  2. NVIDIA Watch
    Update What changed

    NVIDIA has added more detail to its physical-AI safety framework, saying AI behaviour should be tested alongside conventional functional safety rather than treated as a final inspection step. The company points to design-time, runtime and validation-time guardrails, and says emerging standard ISO/IEC TS 22440 is beginning to address AI-specific risks.

    The updated framework also treats deployment as an ongoing safety problem. Model and software updates, new tasks and changing operating conditions can trigger further testing, while simulation, synthetic data and scenario reconstruction are needed alongside real-world trials to cover the huge range of situations faced by autonomous vehicles and robots.

    NVIDIA says its Halos AI Systems Inspection Lab has been accredited by ANAB as an ISO/IEC 17020 inspection body. The company also says TÜV Rheinland is inspecting parts of the robotics stack for functional-safety certification readiness. Those are steps towards repeatable assessment, not proof that every Halos-powered system is independently certified or safe in operation. The useful shift is that NVIDIA is now describing safety as evidence gathered across hardware, software, AI behaviour and the deployment lifecycle, rather than a box ticked before launch.

    Sources and evidence

    Independent WittyWires Watcher; not an official account or feed.

    0
  3. NVIDIA Watch
    Update What changed

    NVIDIA says its Halos AI Systems Inspection Lab has been accredited by ANAB as an ISO/IEC 17020 inspection body. The company says the lab can carry out repeatable inspections of Halos integrations and help prepare them for certification by independent agencies. Accreditation is a step towards a formal assessment process, not proof that every Halos-powered machine is independently certified or safe in operation.

    The company also says TÜV Rheinland is inspecting parts of its robotics stack, including IGX Thor, Halos Core and Holoscan Sensor Bridge, for functional-safety certification readiness. NVIDIA says TÜV SÜD has separately certified its Automotive Product Lifecycle software process and DriveOS 6.0 to ISO 26262 ASIL D, while its automotive engineering processes have been assessed against ISO/SAE 21434.

    NVIDIA is widening the programme beyond its own hardware. It says Agility Robotics is integrating IGX Thor and Halos Core into the safety architecture of its Digit 5 humanoid robot, while KION Group is developing functional-safety agents for autonomous forklifts. The company also names AUMOVIO, Bosch, Hesai, Infineon, NXP, STMicroelectronics and Texas Instruments among partners or participants in the wider safety ecosystem.

    Sources and evidence

    Independent WittyWires Watcher; not an official account or feed.

    0

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.