Author: Humanoid Analytics

Research position paper. Author: Rinat Mirzaitov. Abstract This paper argues that artificial general intelligence cannot be achieved without general-purpose physical AI. The claim concerns an integrated system’s capacity to learn and perform physical tasks, rather than a requirement for a humanoid body or a particular training architecture. It also proposes that humanoid robots could become the primary interface for this intelligence in environments built for people. If general intelligence includes adapting to the environments in which human goals arise, it must cover physical interaction as well as digital reasoning. A system that cannot connect its plans to perception, action and…

Agility’s Digit has the clearest evidence of repeat commercial demand among these three robots. On October 26, GXO and Agility announced plans to expand to two additional warehouses after working together at Flowery Branch, Georgia. The new deployments are expected in early 2027, so they cannot yet be counted as operating sites. Digit: the customer wants more GXO confirms that Digit has moved more than 100,000 totes at Flowery Branch under the companies’ commercial relationship. The task is specific: moving totes and placing them onto conveyors alongside existing automation. The planned additions in Statesville, North Carolina, and Lebanon, Tennessee, strengthen…

Evidence cutoff: 30 September 2026, following the supplied month-end rule. Research completed: 16 October 2026. [1] Figure leads this public-evidence comparison, while Agility’s Digit has the stronger commercial operating record. Those are different achievements. Figure’s advantage comes from measured work in unfamiliar homes. Digit’s comes from customer operations that have moved beyond an initial demonstration. Neither has published the combination of breadth, independence from human help, reliability and economics needed to establish general-purpose physical AI. [1,2,3,4] Applying the supplied Humanoid Leaderboard methodology produces a leading score of 44 for Figure and 38 for Digit. These are calculations from the evidence…

Humanoid robots have found one of their first meaningful commercial markets in an unexpected place: entertainment. Before humanoids can reliably replace workers across factories, warehouses, retail stores or homes, they can already create value by doing something much simpler. They attract attention. That distinction has commercial consequences. Counterpoint Research estimates that global humanoid robot sales revenue exceeded $500 million in 2025. Entertainment and performance accounted for 26% of that revenue, the largest application category in its analysis, ahead of data production at 22%, intelligent manufacturing at 17%, services and guidance at 15%, and warehousing and logistics at 6%. Mathematically, that…

The race to build the best humanoid hand no longer has an obvious leader. Inspire Robots has the strongest disclosed shipment evidence. LinkerBot is pushing high-degree-of-freedom hands toward higher production volumes. Sharpa has concentrated on dense tactile sensing and an integrated manipulation stack. Unitree is combining dexterous hands with its own widely available humanoid platforms. Figure AI is building sensing directly around its Helix AI system. Tesla remains one of the industry’s most closely watched hand programs, but public evidence around its latest hardware is unusually limited. That makes “leadership” dependent on what is being measured. For investors and corporate…

Trajectories, infrastructure, and the safety and security imperative. Sep 25, 2026 – Author: Rinat Mirzaitov Abstract This paper examines the race to build general-purpose physical AI, meaning learned systems that can do open-ended physical work in unseen places and on different robot bodies. Using filings, research, standards and market data available in September 2026, it proposes a six-level scale of physical generality, P0 to P5, and places the field at the boundary between P2 and P3. The best systems reach 56% zero-shot success on household tasks in unseen homes, and about 19,100 humanoids shipped worldwide in the first half of…

Tesla’s Optimus program has never lacked ambitious production targets. What has been harder to see is the manufacturing machinery needed to turn those targets into robots. That may now be starting to change. Chinese outlet The Paper reported on September 21 that Tesla’s robotics team had begun factory audits at suppliers including Tuopu Group, Sanhua Intelligent Controls and Joyson Electronics. The companies are associated with components such as joint modules, actuators and precision structural parts. An unnamed supply-chain source went further, claiming that the suppliers had already received Tesla humanoid-robot orders and that production could follow once quality and compliance…

China does not control 90% of the entire humanoid robotics supply chain. It does control more than 90% of several critical upstream layers that sit inside the robot. That is the more important conclusion. Humanoid Analytics’ current Company Tracker contains 360 organizations across humanoid OEMs, component companies, manipulation specialists, sensing companies, corporate programs and other parts of the ecosystem. China accounts for 221 of those entries, or 61.4%. The concentration rises when the analysis is restricted to the 225 companies and programs classified by Humanoid Analytics (HA) as Core. China accounts for 144, exactly 64.0%. Among 204 tracked Humanoid OEMs,…

Humanoid robotics is developing a training-data market, but it is not yet a commodity market. Robot developers are building dedicated collection systems, specialist companies are offering physical AI data services, open datasets are distributing robot trajectories at scale, and early marketplaces are testing whether robotics datasets can be verified, licensed and sold. The demand is real. The less certain question is where durable commercial value will sit. That question matters for another reason. Data is not simply an input purchased by robot developers. For learned robotic systems, it is part of the path toward greater autonomy. A humanoid operating outside…

XPENG has moved its IRON humanoid program from prototype development toward line manufacturing. On September 7, the company said its humanoid robot production lines had been commissioned and that an IRON robot completed production and walked off the line autonomously. XPENG also said more than 80% of the lines’ core processes are automated. For investors, this is a meaningful manufacturing signal because it shows physical production infrastructure rather than another future capacity target. The main limitation is source quality: the evidence is company-controlled, and no independent source reviewed by Humanoid Analytics establishes sustained output, external customer delivery or real customer…