What supports the claim?
We separate vendor statements, media reports, customer signals, operating metrics, and corrections.
Methodology
How we turn public evidence into market intelligence.
Humanoids Watch evaluates humanoid robots through public evidence, deployment readiness, use-case fit, commercial posture, and buyer risk. Scores are signals for diligence, not certification or endorsement.
Evaluation Areas
What every assessment is trying to answer.
Could a buyer evaluate it?
Deployment readiness reflects public signals around access, workflow fit, support model, and maturity.
Where does it make sense?
Robots and deployments are interpreted through concrete work, not generic humanoid capability.
What is still missing?
Unknown uptime, intervention rate, autonomy boundary, economics, and safety posture remain visible.
Score System
Scores are compact signals, not final answers.
Robot and deployment pages use scores and evidence labels to make comparison easier. The written summary, linked deployment records, and visible limitations matter as much as the number.
Deployment Readiness
0-100The main buyer-facing score. It estimates whether a robot appears suitable for evaluation in a defined use case.
Reality Score
0-100A substantiation score for the robot itself: hardware maturity, demonstrated capability, and public proof.
Evidence Quality
Weak to Very StrongA source-strength label that shows whether the public record is thin, moderate, customer-confirmed, or metric-backed.
Public Claims
Not every announcement is evidence.
Demos, partnerships, and vendor announcements can matter. They become stronger when they reveal who is involved, what work is being done, when it happened, and what remains unproven.
Demo
Useful as a capability signal, but not treated as deployment evidence.
Partnership
Stronger when the customer, task, and commercial context are named.
Pilot
Buyer-relevant when the workflow, site, duration, or customer confirmation is public.
Operating metric
Strongest when uptime, intervention rate, throughput, safety, or economics are disclosed.
Risk And Missing Proof
Uncertainty stays visible.
The absence of public proof is not always negative, but it is buyer-relevant. Profiles should show what is known and what still needs diligence.
Update Process
Assessments change when the evidence changes.
New signal
A public source, correction, deployment record, or vendor update appears.
Review
We check source quality, date, specificity, and relevance to robot, vendor, deployment, or use-case pages.
Profile update
Scores, summaries, linked records, red flags, or review language change only when the evidence justifies it.
Use The Methodology
Apply it where the data lives.
The methodology is reflected across robot profiles, vendor profiles, deployment records, and use-case pages.