Skip to content
Blog
Deep dives8 min

A $13,000 robot vs a $900-a-month waiter

Robot waiter vs a hired one: the full three-year cost — service, breakdowns, taxes, depreciation. We count the whole table, with a source for every number.

Compare like with like. A robot waiter does not replace a waiter — it replaces the runner function: the legs between kitchen and floor. On a busy floor that is two shifts of running, so the right column holds two wages with taxes, not one. The left column is not the FOB price but the full cost of ownership: the machine landed and commissioned, plus the Standard service plan for the whole horizon.

NXM-ROI / 36 monthsRobot vs 2 runner wages
Robot, turnkey (full landed cost)$17,800
Standard service · 3 years$4,800
Robot total over 36 mo$22,600
Runner: $900 × 1.3 tax × 2 shifts × 36$84,240
Difference over 3 years≈ $61,600 in the machine's favour · payback ~11 mo

The 1.3 tax coefficient is conservative for the CIS. 85% runner-task replacement is already built into the package calculator.

The table omits the lines both sides of the argument usually hide. For people: hiring and training with quarterly turnover, sick days, no-shows on Saturdays. For the machine: consumables, and the fact that it brings no tips. Across five sites we checked, these roughly cancel out.

The robot is not cheaper than a waiter. It is cheaper than waiter turnover.

Part 02

Three cases where the maths breaks

Under 60 deliveries a day

The robot has nothing to carry — it pays back on flow, not presence. A small floor is more honest with people.

Stairs, high thresholds, narrow aisles

Every detour eats delivery minutes. Sometimes it is cheaper to re-lay a threshold; sometimes, not to buy the machine.

Service “via China”

A week of downtime waiting for a sea-freight part turns an 11-month payback into 18. Price service in, not separately.

Part 03

Where $17,800 comes from: the full budget

The “$13,000 robot” in the headline is the FOB price (free on board — the cost at the Chinese port). Comparing it to a wage is the sellers' favourite trick: between the port and your floor lies another third of the cost. Here is the full budget of the very machine from the case — line by line, no “to be confirmed later”.

Full turnkey budget, a typical CIS project
Budget lineAmountComment
Machine, FOB Shenzhen$12,900the advertised “robot price”
Freight and insurance$1,450consolidated container, DG battery paperwork
Duty and VAT$2,050depends on HS code and country
Call stations and dock$78016 stations, charging dock, mounts
Launch: mapping, setup, training$6202 engineer-days on site
Turnkey total$17,800+37% over the advertised price

The Standard service plan ($133/mo) is not in the launch budget — it is a separate ownership line.

Part 04

Sensitivity: if your numbers differ

Any ROI calculation is only honest together with its sensitivity: move one variable and the conclusions may flip. We ran three typical deviations from the base scenario through the same formulas.

Why a 36-month horizon: it is a conservative estimate of the machine's life before the first major replacement — the battery. Manufacturers claim five to seven years for the platform; we calculate on three, so the remaining lifespan stays a safety margin of the maths rather than its precondition. If the machine lives longer — and the telemetry from our sites says it will — every following shift runs at pure profit.

Same machine, different inputs — 36-month horizon
ScenarioDifference over 3 yearsPayback
Base: 2 shifts, $900 wage≈ $61,600~11 mo
$600 wage — regional city≈ $33,600~13 mo
One runner shift, not two≈ $19,500~18 mo
Two machines on a large floor≈ $125,000~8 mo

Service and consumables included in all scenarios. Below 60 deliveries a day the calculation stops working — see part 02.

What is deliberately left out

The calculation contains no “wow effect” and no rise in average ticket — both exist, but we cannot promise them. Nor any residual value of the machine: the second-hand market for service robots in the CIS is only forming. Both corrections favour the machine, so the calculation is conservative — we would rather you not be disappointed than buy faster.

How to measure your “before”

The main error in do-it-yourself maths is eyeballing the “before”. A week of measurement before the decision saves months of disappointment: count deliveries per shift from receipts or the KDS, stopwatch three or four kitchen-to-table runs at peak, log the person-hours spent running. Those three numbers are the calculator's inputs. And if the measured flow comes in under 60 deliveries a day — you have just saved $17,800 for free.

Takeaway

Run it on your numbers — the formulas are open

This whole calculation is built into the Restaurant package calculator: move the wage, the shifts and the number of machines — and get your payback under the same assumptions as this article. Attach the result to your request and the conversation with an engineer starts from your numbers.

Your floor — your numbers

Two calculator scenarios: a restaurant floor and hotel room delivery. Same formulas as this article.

Author · Nexum analytics · DXB
Next step

Want to map this onto your site?

The two-minute audit suggests the right format — catalog, package or custom project. Or write to us directly: the conversation starts with your numbers, not a slide deck.