The latest Humanoid Deployment Tracker shows a sharper and more cautious picture of humanoid robotics deployment progress: out of 164 companies tracked, only two currently receive a best evidence score of 10. Agility Robotics and Figure AI are the only companies in the tracker classified at the highest level, which Humanoid Analytics labels as “operating proof with metric.” That matters because the humanoid robotics market is crowded with prototypes, pilots, preorders, factory demos, and production claims, but very few companies have public evidence tied to named customers, real operating environments, and measurable work.
This is not the same as saying only two humanoid companies matter. It means only two companies currently clear the tracker’s highest evidence bar. The rest of the market remains earlier, less verified, or more dependent on announcements, company-sourced claims, shipment claims, preorders, and future deployment plans.
The distinction is important. In humanoid robotics, a company can be well funded, technically credible, and strategically important without having strong deployment proof. A customer logo does not prove useful work. A product launch does not prove customer adoption. A factory video does not prove repeat operations. The tracker’s value is that it separates those signals instead of treating them as equivalent.
The Companies At Score 10
Agility Robotics receives a best evidence score of 10 for Digit. The tracker lists five event rows and identifies GXO Logistics, SPANX, Toyota Motor Manufacturing Canada, Mercado Libre, and Schaeffler Group as named customers or partners. Its best deployment type is classified as commercial deployment, with active status. That makes Agility one of the clearest public leaders in humanoid deployment evidence, especially because its signal is not limited to a single announced relationship.
Figure AI also receives a best evidence score of 10, based on Figure 02 and Figure 03 evidence tied to BMW Group Plant Spartanburg. The tracker lists two event rows and classifies Figure’s best deployment type as commercial or operational customer deployment, with completed deployment status. Figure’s evidence is strong because it is linked to a real industrial customer site, but the tracker also notes that Figure’s external commercial deployment remains less proven than its capital position and manufacturing ambitions.
The top of the tracker is narrow. Ranked by best evidence score, only Agility Robotics and Figure AI reach the maximum score, while UBTECH Robotics sits in the next tier below them.
| Rank | Company | Evidence Score |
|---|---|---|
| 1 | Agility Robotics | 10 |
| 1 | Figure AI | 10 |
| 3 | UBTECH Robotics | 8 |
| 4 | Apptronik | 7 |
| 4 | Humanoid (UK) | 7 |
| 6 | AgiBot | 6 |
| 6 | Sanctuary AI | 6 |
| 8 | Tesla | 5 |
| 9 | 1X Technologies | 4 |
| 9 | Fourier Intelligence | 4 |
| 11 | NEURA Robotics | 3 |
| 11 | Pudu Robotics | 3 |
UBTECH Robotics sits just below the top tier with a score of 8. That is still a strong signal. The tracker lists 10 event rows and named customers or partners including ZEEKR, Geely, BYD, NIO, Dongfeng Liuzhou Motor, FAW-Volkswagen Qingdao, SANY Renewable Energy, and Foxconn New Energy Vehicle R&D Centre. But the classification remains “named customer deployment / agreement,” not “operating proof with metric.” In practical terms, UBTECH appears to have meaningful industrial traction in China, but the market still needs more customer-by-customer verification, operating data, robot counts, and evidence of sustained use.
Why The Gap Matters
The score distribution is the real story. Most companies in the tracker do not have verified event rows or strong public deployment proof. Many may still be relevant, technically capable, or commercially interesting, but they have not yet reached the same evidence level as Agility Robotics and Figure AI.
Below the top tier, the evidence becomes much more conditional.
Apptronik scores 7, with named relationships involving Mercedes-Benz, GXO Logistics, and Jabil. That is commercially meaningful, but the tracker classifies it as a named pilot or deployment claim rather than operating proof. Humanoid also scores 7, based on a future rollout agreement with Schaeffler. That makes it an important European watch item, but its status remains planned.
AgiBot and Sanctuary AI score 6. AgiBot has factory livestream and production-line trial signals, but its large shipment claims still need to be separated from customer deployment proof. Sanctuary AI remains relevant because of Phoenix, dexterity, and AI-control work, but the tracker notes that recent independently verified deployment evidence is limited.
Tesla scores 5. That reflects strategic importance and internal production-use claims, not strong external customer deployment proof. 1X Technologies and Fourier Intelligence score 4, reflecting commercial signals with weak deployment proof. NEURA Robotics and Pudu Robotics score 3, which the tracker classifies as demo or product signal only.
This should temper the market narrative. Humanoid robotics has moved beyond pure lab demonstrations, but it has not yet reached broad commercial deployment. The strongest evidence is still concentrated in logistics, manufacturing, and industrial environments. That is not surprising. These settings have repetitive workflows, known sites, enterprise customers, and clearer automation budgets. They are more plausible early markets than open-ended home or public-service use.
What The Tracker Proves, And What It Does Not
The tracker proves that there is a small group of companies with stronger deployment evidence than the rest of the field. It also shows that Agility Robotics and Figure AI currently occupy the top evidence tier. UBTECH deserves close attention because of the volume of named industrial activity, but its claims still need more operating metrics and independent customer confirmation.
The tracker does not prove that humanoid robots are commercially mature. It does not prove positive unit economics, high uptime, low intervention rates, safety performance, repeat orders, or scalable manufacturing. It also does not prove that early pilots will turn into large deployments.
The next evidence to watch is straightforward: more named customer confirmations, robot counts, deployment duration, paid commercial terms, operating hours, task success rates, uptime, safety records, intervention rates, repeat deployments, and customer expansion. Those are the signals that would move the market from interest to adoption.
For now, the Humanoid Deployment Tracker points to a market with two clear top-tier deployment-evidence leaders, one strong China industrial contender, and a long tail of companies still waiting for proof. That is not a weak market. It is an early market. The difference matters.
Sources:
Humanoid Analytics, “Humanoid Deployment Tracker”: https://humanoidanalytics.com/humanoid-deployment-tracker/