Catalog Hygiene Compared: What 4,435 Robot Listings Say About Dealer Prices

We have spent six weeks reading robot dealer catalogues for prices. This piece looks at the catalogues themselves. On 12 August 2026 I pulled the complete public product data from eight US storefronts that sell robots, drones or ROVs — 3,109 products and 4,435 purchasable variants — and counted the things that make a price hard to trust: strike-through prices, placeholder prices, duplicate listings and dead stock.

The stores: five drone dealers (Maverick Drone, Infinit Drones, DSLRPros, DroneFly, Volatus), one robot specialist (US Robot Store), one manufacturer store (Rover Robotics) and one consumer brand store (HOBOT US). All prices in USD, confirmed through each store's own cart endpoint.

The scoreboard

Discounting is a store habit, not a market rate

The share of listings carrying a struck-through "was" price runs from 1.0% to 40.7% across these eight stores on the same day. Widen it with catalogues we have read before and the range gets absurd: 94.6% of variants at a pipe-inspection dealer in our 3 August read, versus one single product out of 302 variants at an exoskeleton specialist four days later.

Nothing about robots changed between those two catalogues. The merchandising did. If you are inferring "the market is discounting" from strike-throughs, you are measuring which platform the dealer configured, not what anyone is willing to pay.

The was-price that is not a history

The clearest case sits at US Robot Store, where 176 of 394 variants carry a compare-at value that cannot be a former price:

• Seven Unitree G1 humanoid configurations priced $53,900 to $73,900 all show "was $20,900" — the same number on all seven, and *below* every one of them. $20,900 is roughly the entry G1, not the history of a $73,900 machine.
• A Go2 industrial quadruped at $13,200 shows "was $3,500".
• Twenty-one listings — including a $100,000 quadruped and an $11,800 robot arm — show "was $0.00".
• 129 variants show a was-price exactly equal to the asking price: a 0% discount rendered as a discount.

This is the fifth species in the catalogue of fake prices we have been keeping: after four machines sharing one price, $0.01, $0.00, and $17.81 for a six-month software licence, here is *the sibling model's price parked in the discount field*. Volatus and Rover carry the same defect at smaller scale — Rover's MEGA charging dock is $1,450 marked down from $950.

$0.00 means "call us"

Infinit Drones lists 111 variants at $0.00, and they are almost all services: BVLOS training, counter-UAS awareness, search-and-rescue courses, stockpile mapping. Maverick does the same thing inside a single product — its Fleet Maintenance Program prices the Essential and Priority tiers ($2,499 to $11,499) and leaves the "Custom Fleet" tier at $0.00.

That is defensible as merchandising and destructive as data. Every price aggregator, comparison site and AI answer engine reading these feeds sees a $0 robot service. We traced where invented robot prices come from two weeks ago; this is the raw material.

The same price on twenty-one different things

At Maverick, 51 distinct prices above $500 are each shared by four or more unrelated products. The largest cluster is $999.00 across 21 products — parachute systems, an agricultural spreader, a 150 W flood light, a wheeled hard case. DSLRPros has $999.00 across seven, and $1,643.00 across five gas-sensing modules that differ only in which gas they detect.

Some of that is real (round-number pricing clusters naturally). Some is the signature we first named on a service-robot dealer, where four machines with completely different payloads all showed $20,580.00. The test is whether the products in a cluster have anything to do with each other.

Empty shelves priced like full ones

Volatus lists 511 variants and 350 of them are out of stock (68.5%) — including every Vantage Robotics SKU it carries, at prices identical to the dealers that have them. That is not dishonest; it is just not a quote. The archive version of this is worse: a research-arm maker's store carried 2,794 product records of which only 591 had any purchasable variant, with the legacy arms still listed at prices no one can transact.

Duplicates: the benign kind and the expensive kind

DSLRPros lists the Vantage Trace Public Safety bundle twice — once under its own house title — both at $12,599. Harmless. Compare the pipe-inspection dealer where listings literally titled "(Copy)" were priced up to $5,000 above the originals, and where one machine appeared under three different was-prices. Duplication only costs you money when the copies disagree.

How I would read a robot catalogue

1. Check whether the was-price is possible. If it is below the asking price, equal to it, or identical across configurations, the store's discount field is decorative — and so is the discount.
2. Count the cluster. Search the store for the exact price you were quoted. If four unrelated machines share it, you have a placeholder.
3. Check stock before you compare. Two-thirds of one catalogue here is unbuyable; its prices are the cleanest-looking in this article.
4. Read the buy button, not the number. Rover Robotics publishes exact prices in its data and shows "Request a Quote" on every product page — we priced the line anyway, with that caveat attached.

None of these four checks takes longer than a minute, and all of them are things a buyer can do that a price aggregator will not.

Sources

• Maverick Drone, Infinit Drones, DSLRPros, DroneFly, Volatus Drones, US Robot Store, Rover Robotics and HOBOT US public product feeds and cart endpoints, all read 12 August 2026
• GrabaRobot, Pipe Inspection Crawler Robot Price Guide, 3 August 2026
• GrabaRobot, Industrial Exoskeleton Price Guide, 7 August 2026
• GrabaRobot, The Answer Layer Is Inventing Robot Prices, 2 August 2026

*Limitations: single-day snapshots; "broken price markers" is my classification of compare-at values that are zero, equal to or below the asking price, plus prices of $0.00/$0.01/$1.00; duplicate detection is by normalised title, so re-branded duplicates are undercounted. Store size differs by an order of magnitude, so percentages, not counts, are the comparable figures.*