Residual-value data for the businesses underwriting humanoid robots.
Lenders, lessors, insurers, dealers, and analysts are pricing risk on humanoid robots and robot dogs — assets that can shed ~40% of their value in a year, with almost no public resale market to anchor to. Valumech licenses the underlying dataset: verified new prices, availability, and modeled residual curves for 104 models, so your residual and pricing assumptions rest on sourced numbers instead of guesses.
What you get
- 104 models across humanoids, robot dogs, and home and service robots — the notable machines actually on the market, not a padded count.
- Verified new prices, each with the source URL and the date it was last checked. 33 of the 104 models carry a genuine public price; the rest are honestly marked unpriced rather than filled with a guess.
- Availability status per model — shipping, pre-order, or announced — so you know which assets actually exist in the world.
- Modeled residual curves, clearly labeled as modeled: a transparent tier-based depreciation model, not resale-backed appraisals, because the used market for these robots barely exists yet.
- Price-change history as it accrues — prices are re-checked daily, so the dataset builds a real record of how new prices move over time.
A real sample
Two rows from the dataset, in the shape you'd license (JSON). Every price carries its source_url and verified_date; the residual_note field is model output, marked MODELED — never a verified transaction.
[
{
"id": "unitree-g1",
"maker": "Unitree Robotics",
"name": "G1",
"category": "humanoid",
"price": {
"amount": 13500,
"currency": "USD",
"source_url": "https://shop.unitree.com/products/unitree-g1",
"verified_date": "2026-07-15"
},
"status": "shipping",
"residual_note": "MODELED — ~40%/yr geometric curve, ~60% retained at 12mo; not resale-backed (no verified used comps yet)"
},
{
"id": "deep-lynx-m20",
"maker": "DEEP Robotics",
"name": "Lynx M20",
"category": "robot-dog",
"price": {
"amount": 61200,
"currency": "USD",
"source_url": "https://shop.deeprobotics.us/products/lynx-m20",
"verified_date": "2026-07-15"
},
"status": "shipping",
"residual_note": "MODELED — ~30%/yr geometric curve, ~70% retained at 12mo; not resale-backed (no verified used comps yet)"
}
]Prices shown are real, publicly sourced manufacturer figures. Residual fields are modeled estimates from our published depreciation method, not resale comps.
Data dictionary
What each field in the sample above holds, spelled out plainly. The full licensed export also includes specs (height, weight, DoF, debut year), left out of the 2-row demo for brevity. Only residual_note is model output; every other field is a real fact or left blank.
| Field | Type | What it holds |
|---|---|---|
id | string | Stable slug, the primary key used across every export and URL. |
maker | string | Manufacturer name, e.g. Unitree Robotics. |
name | string | Model name as sold, e.g. G1. |
category | enum | Semantic market category: humanoid, robot-dog, home-robot, mobile-manipulator, or service-robot, assigned from what the robot actually does rather than its form factor. |
status | enum | Availability: shipping, preorder, announced, or discontinued. |
price.amount | number | null | Base-config list price. Null when no manufacturer or retailer has published one, never an estimate. |
price.currency | string | null | ISO currency code for price.amount. |
price.source_url | string | null | The manufacturer or retailer page the price was checked against. |
price.verified_date | date | null | Date price.amount was last confirmed against price.source_url. |
specs | object | Height, weight, degrees of freedom, and debut year. Each is tied to a spec source when the maker publishes it; unverified figures are left blank rather than guessed. |
residual_note | string | A tier-based depreciation estimate built from our published method. Not a resale-backed appraisal, since a verified used market for these robots barely exists yet.Modeled |
Coverage right now
Every figure below is computed from the live dataset at build time — none of it is typed in by hand, and it updates as the dataset grows.
Source-URL coverage counts models carrying a spec source plus configs carrying a price source, against the total number of models and configs in the dataset.
What it's for
Residual underwriting
Set loan-to-value and end-of-term residuals on robot-backed financing, with sourced new prices and a transparent depreciation curve behind every number.
Lease & insurance pricing
Price monthly lease rates, replacement cost, and total-loss values from current verified prices instead of a vendor's quote or a stale spreadsheet.
Competitive intelligence
Track where the market is moving — who launched, who cut price, and how fast — across every notable model, with provenance you can check.
Pilot pricing available.
Most pilots start as a scoped monthly export — the slice of models and fields you actually need. Tell us the use case below and we'll shape a pilot around it.
How the numbers are built
Every figure is sourced or left blank — never guessed. New prices come from manufacturer pages and official channels with full provenance; residual estimates come from a transparent, tier-based model we label as modeled. Read the full methodology and our editorial policy before you build on the data.
Request pilot access
Tell us who you are and what you'd use the data for. We read every request — expect a reply within a few business days.