A monthly, evidence-graded read on humanoid progress.
Each month we gather deployments, hiring, field video and vendor claims, check every source, and score every robot on the same scale — so progress is visible, and so is the distance still to go.
How a score is built.
Every robot is read against the same evidence, graded on the same ladder and published on the same day of the month. Early stages are not a verdict — they are simply where the public record stands today.
Every claim is placed on a six-stage evidence ladder. A score can rise only as far as the public record has travelled.
Customer or site named.
Evidence required · Publicly attributable customer or site.
- Scores are capped by the evidence stage
- Every claim carries its source
- Scoring is not open to vendor input
The published cohort.
Every shape below is the published score — the same numbers the Index ranks on, across eight axes and forty sub-axes. Filled is capability; the dashed outline is how much of it the public record supports.
UBTech Walker S
- Rank
- #1
- Capability score
- 1.70
- Proof score
- 0.13
- Coverage
- 29/40
- 0.19→1.60+1.41
- 0.15→1.80+1.65
- 0.09→1.60+1.51
- 0.05→2.00+1.95
- 0.05→0.40+0.35
- 0.32→2.80+2.48
- 0.08→2.60+2.52
- 0.11→0.80+0.69
Deployments, hiring and field video.
Named customer pilots, job posts, and on-camera footage each carry different weight. Every item is labelled with what it evidences.
Hiring intel
Open jobs →Hiring is a commercialization signal, not capability proof.
Send evidence for review.
Vendors and operators share deployment proof, certifications and customer references with us confidentially. We verify it against the source, grade it on the evidence ladder, and publish only what the record supports.
What you send stays private. Only the graded claim and its public source ever appear on the site.
Get the monthly briefing.
The deployments, hiring moves, and Index changes from the past month. One email a month. Unsubscribe in one click.