Topic

robotics deployment

10 signals

Economics3 hours ago

Software becoming the biggest bottleneck to physical AI innovation, finds QNX research

QNX, the BlackBerry-owned embedded OS division, released an OnePoll survey of 1,000 robotics software developers finding 27% cite software architecture and integration as their biggest performance bottleneck versus 16% for hardware, while 91% still run real-time workloads at least partly on general-purpose operating systems. Two-thirds of respondents reported certification-related project delays, rising to 70% in the UK and Germany versus 56% in China.

Why it matters

The number to dwell on is the certification delay gap: 70% of UK and German teams report project slippage from certification versus 56% in China. That is a 14-point speed advantage that compounds across product cycles, and it is showing up in the survey data before it shows up in deployment counts. For anyone modeling time-to-revenue on a humanoid or mobile manipulator, the regulatory cost of operating in Europe is now quantifiable and material.

The other tell is that 91% of developers admit they are running deterministic, safety-critical workloads on general-purpose operating systems (read: Linux), and 86% of that group say they are open to switching. That is a vendor's dream survey, and QNX commissioned it, so discount accordingly. But the underlying tension is real: the robotics stack was built on ROS plus Ubuntu because it was fast to prototype, and nobody wants to be the company that ships a hospital robot certified on a kernel they cannot freeze. Expect a wave of safety-certified RTOS procurement over the next two years, with QNX, Wind River VxWorks, and Green Hills INTEGRITY as the obvious beneficiaries.

InfrastructureMay 28

NVIDIA Research Advances Robotics From Simulation to the Real World

Source: NVIDIA Blog

NVIDIA Research presented 28 papers at ICRA, with eight focused on simulation-to-real transfer across multi-arm scheduling (ScheduleStream, 3x speedup), cross-embodiment navigation (COMPASS, ~80% success on real robots), grasping in clutter (Grasp-MPC, 75% vs 41% baseline), precision assembly (SPARR, 38% success improvement; Refinery, 91% sim success), and vision-language-action grounding (PEEK, up to 41x accuracy gain; SEAL, 15% gains). NVIDIA also said its Physical AI Dataset has passed 15 million downloads, and roughly 50 ICRA papers from CMU, ETH Zurich, MIT and UT Austin cite NVIDIA-accelerated simulation or compute.

Why it matters

Strip away the paper-by-paper benchmark recitation and the operator read is straightforward: NVIDIA is locking in Isaac Lab, Isaac Sim and the Physical AI Dataset as the default substrate for robot policy training, the same way CUDA locked in GPU compute. Fifty ICRA papers from CMU, ETH, MIT and UT Austin referencing NVIDIA-accelerated stacks is the leading indicator. The next generation of robotics hires arrives already fluent in Isaac, which is how a tools moat compounds.

The technical numbers worth tracking are the sim-to-real deltas, not the absolute success rates. Grasp-MPC going from 41% to 75% on novel cluttered grasps, SPARR cutting cycle time 30% on NIST assembly, PEEK delivering a 41x accuracy multiplier on sim-trained policies: these are the levers that move warehouse and assembly economics from "demo" to "billable." For operators evaluating whether to in-house policy training or buy from a foundation-model vendor like Physical Intelligence or Skild, the question is whether these techniques ship as reusable primitives in Isaac or stay locked in research repos. The GitHub release for ScheduleStream suggests the former, which would further pull policy R&D onto NVIDIA's rails.

LaunchesMay 28

Robot dogs help police catch car break-in suspects at Atlanta apartment complex

Source: Robotics & Automation News

Undaunted, an Atlanta-based robotic patrol startup, said its security robot dogs tracked and helped corner two suspects accused of breaking into vehicles at the Columbia Crest Apartments in West Midtown on May 21, with one robot following the suspects while another alerted Atlanta Police. The company says it has deployed robotic systems at more than three dozen properties across the Atlanta region in the past year and is evaluating expansion into additional markets.

Why it matters

The interesting datapoint isn't the arrest, it's the customer profile. Undaunted is selling robot patrols into affordable housing, where the buyer (Columbia Residential) is underwriting against vacancy losses driven by perceived safety, not against a security guard wage line. That changes the unit economics math entirely. A Knightscope or Boston Dynamics Spot deployment that has to beat a $20/hour guard contract is a hard sell at three dozen properties; one that has to move occupancy by 200 basis points on a sub-performing affordable asset is a much easier ROI conversation for an asset manager.

The operational tell is the three-robot density on a single property with autonomous handoff (one pursues, one dispatches). That implies Undaunted is running fleet coordination software, not just selling hardware as a guard replacement. For operators watching the security robotics segment, the question is whether Undaunted owns the platform or is reselling Unitree or Deep Robotics quadrupeds with a patrol-management layer on top. The Atlanta footprint is small but it's the first multifamily deployment with a public incident log, which is what insurance carriers and REIT asset managers will actually price against.

LaunchesMay 27

Figure partners with Catalyst Brands to deploy humanoid robots in logistics operations

Source: Robotics & Automation News

Figure signed a commercial agreement with Catalyst Brands to deploy humanoid robots across the retailer's distribution network, starting at a distribution center in Reno, Nevada. Catalyst Brands is the holding company for JCPenney, Aéropostale, and Brooks Brothers.

Why it matters

Catalyst is Figure's first publicly named retail logistics customer outside the BMW manufacturing pilot, and the choice of site matters more than the brand list. Reno is the standard western US e-commerce fulfillment node (Amazon, Walmart, Thredup all run hubs there), which means whatever throughput Figure proves at JCPenney's DC translates directly into pitches for the next twenty contracts. Apparel distribution is also the easier end of the humanoid manipulation problem: soft goods, polybags, totes, not mixed-case grocery.

The number to watch is units on the floor, not the press release. Figure claims one robot per hour off the BotQ line and 350+ Figure 03s built to date, so Catalyst could plausibly absorb 50 to 100 units across the network without straining supply. If the Reno deployment stays at single-digit robots through 2026, this is a data-collection deal dressed as a commercial one (Helix needs the operational hours). If it scales to a full shift replacement at one DC, that is the first real humanoid unit-economics datapoint in US retail logistics.

LaunchesMay 27

Figure ramps up humanoid robot manufacturing at unprecedented speed

Source: Robotics & Automation News

Figure AI says its BotQ facility has scaled Figure 03 production from one unit per day to one unit per hour in under 120 days, with more than 350 third-generation robots and over 9,000 actuators across 10+ SKUs built to date. The company reports a 99.3% first-pass yield on its battery line and over 80% first-pass yield on full robots, supported by 150+ networked workstations, 50+ in-process inspection points, and 80+ end-of-line tests.

Why it matters

One unit per hour is roughly 7,000 to 8,000 robots a year at single-shift, which would put Figure ahead of every Western humanoid competitor on installed base if the line holds yield. The number that matters more than throughput is the 80% first-pass yield on full robots and 99.3% on batteries. Most humanoid programs are still rework shops, where every unit gets torn down and re-flashed before it ships. Crossing 80% first-pass is the threshold where bill-of-materials cost stops being theoretical and you can actually quote a customer a price.

The quieter signal is the 9,000 actuators across 10+ SKUs. Figure is vertically integrating the actuator stack rather than buying from Harmonic Drive or Nabtesco, which is the same playbook Tesla ran on motors and the Chinese humanoid OEMs are running off the EV supply chain. For operators evaluating humanoid vendors in 2026, the question is no longer demo quality. It is who can ship 1,000 units against a PO, and who controls their own actuator line when the next bottleneck hits.

ChinaMay 25

China to give every humanoid robot a digital ID in push to boost industry standards

Source: Tech - South China Morning Post

China's Ministry of Industry and Information Technology launched the Humanoid Full Lifecycle Management Service Platform on Friday, assigning every domestically manufactured humanoid robot a unique digital ID to track it from production through recycling. The program is led by the Humanoid Robotics and Embodied Intelligence Standardization (HEIS) committee, with companion lifecycle-management guidelines covering manufacturers, service providers, sellers, users, and recyclers.

Why it matters

Beijing is building the regulatory plumbing for humanoid scale before the West has agreed on what a humanoid even is. A national serial-number registry that follows a unit from factory floor to scrap line gives MIIT real-time visibility into production volumes, deployment sites, and failure rates across Unitree, Agibot, UBTech, XPeng, and the rest of the cluster. That data asymmetry compounds: regulators can set safety standards, subsidy eligibility, and export rules using ground truth that competitors and foreign buyers will not have.

For operators procuring Chinese humanoids, the ID becomes a de facto compliance hook. Expect insurance underwriting, factory-floor deployment permits, and eventually cross-border shipments to key off it, similar to how VIN registries shaped the auto industry. The quieter implication is on the recycling end: mandating end-of-life tracking signals Beijing already expects fleet turnover at a scale that requires industrial reverse logistics, which is not how anyone is talking about humanoids in the US yet.

LaunchesMay 24

Sortera uses physical AI to double capacity in a Tennessee sorting facility

Source: The Robot Report

Sortera Technologies brought its second AI-driven metal sorting facility, in Lebanon, Tennessee, to full operational status this month, lifting combined annual processing capacity across its Lebanon and Markle, Indiana sites to roughly 240 million pounds (108.8 million kg). The plant uses Sortera's sensor-and-AI sorting platform to convert mixed alloy scrap into higher-purity aluminum feedstock for automotive, construction, and aerospace buyers.

Why it matters

The interesting number is not the 240 million pounds, it's that Lebanon produced sellable high-purity material in its first week and came online on schedule and on budget. Most AI-powered sorting deployments stall in commissioning because the vision and classification models trained on one feedstock mix don't generalize to a new region's scrap stream. Sortera replicating Markle's performance in Tennessee suggests the platform has crossed from bespoke to repeatable, which is the precondition for a third and fourth site.

For operators in domestic metals and EV supply chains, this is a unit-economics story. Recycled aluminum at roughly 95% less energy than virgin production becomes a hedge against power prices and tariff exposure on imported billet, and Sortera is positioning as a regional feedstock vendor rather than a scrap exporter. The question for the next round of capacity additions is offtake structure: are automakers and aerospace primes signing multi-year volume contracts at a premium to LME, or buying spot. That determines whether facility three gets project-financed or has to be equity-funded.

BonusMay 23

The future of physical AI isn’t humanoid; it’s task-specific and cost-efficient

Source: The Robot Report

Hailo's vice president of physical AI, Yaniv Sulkes, argues in a Robot Report contributed piece that task-specific robots running edge AI will scale faster than general-purpose humanoids, citing hardware, dexterity, energy, and cost constraints. He points to deployments like Husqvarna's AI-enabled robotic lawn mowers, which use Hailo edge processors for on-device sense-think-act loops, as the template.

Why it matters

This is a Hailo positioning piece, and the cap-table read matters: Hailo sells edge inference chips and loses if the humanoid narrative consolidates compute onto Nvidia Jetson-class modules or cloud-tethered VLAs. The argument is still directionally right for operators sizing 2026 deployments. The unit volumes already shipping (robotic vacuums, lawn mowers like Husqvarna's, last-mile delivery, ag sprayers) all clear the bar Hailo describes: bounded task, embedded inference, no cloud in the control loop. Humanoid pilots are not in that category yet.

The number to watch is bill-of-materials. A task-specific robot built around a sub-$50 edge accelerator and a fixed-function actuator stack pencils out at consumer or light-commercial price points. A humanoid with 40-plus degrees of freedom and a 100W-class compute module does not, regardless of how good the foundation model gets. Operators evaluating physical AI vendors should ask where inference runs, what the per-unit silicon cost is, and whether the control loop survives a dead network. Those three questions separate shipping product from demo reels.

DealsMay 22

GE Vernova to acquire Robotech Automation to expand robotics integration

Source: The Robot Report

GE Vernova signed an agreement to acquire Robotech Automation, a 35-person systems integrator based in Longueuil, Quebec, folding it into GE Vernova's Advanced Research Center to deploy robotics across its supply chain.

LaunchesMay 21

Humanoid Secures Bosch and Schaeffler Deals to Scale European Robot Production

Source: The Robot Report

UK-based Humanoid has signed a binding phased deployment and supply agreement with Schaeffler to put its HMND robots into two German factory sites starting December 2026, targeting a four-digit unit count across Schaeffler's global facilities by 2032. The company also partnered with Bosch for manufacturing and distribution in Europe, and separately with Siemens.

← All signals