The compliance gap between a demo and a deployment
The distance from “it works on stage” to “it is legal on your floor”.
Spare parts and the uptime threshold
Below a number, a robot stops being an asset. Where that line sits.
Uptime is the number every quote implies and almost none state — until the machine falls below it.
The threshold nobody quotes
Manufacturers describe reliability in ranges, not guarantees, because uptime is set as much by the parts pipeline behind a robot as by the machine itself. A shift lost to a missing sensor reads the same on a client’s floor as a shift lost to a software fault.
Operators who track this closely treat uptime as a service-contract metric before it is an engineering one. The lever that moves the number is the response window and the parts network behind the machine, not the spec sheet in front of it.
“The robot rarely fails. The queue for the part that fixes it does.”
A parts pipeline is a logistics system wearing a maintenance name.
What a parts pipeline really is
A well-stocked factory warehouse does nothing for a machine standing idle on a client floor if the part that fixes it is a week away by freight. What matters is the distance between the fault and the fix, not the distance between the fault and the factory.
Serious operators map the parts most likely to fail before a machine goes live, and hold a small local stock against them — rather than discovering the gap at the first breakdown.
WHAT TO TAKE AWAY
Ask for the parts and technician network before asking for the uptime number.
A fault fixed on site in hours is not the same guarantee as one fixed at the factory in weeks.
Price the service contract as the lever that moves uptime, not the hardware alone.
Below a certain uptime, a robot stops being an asset and turns into a liability on the floor.
The fix that keeps a robot earning is almost always local, never back at the factory.
Local, not back at the factory
A fault that requires shipping the whole unit home removes it from the floor for weeks, and a client rarely tolerates that twice. The service model that survives is one where a technician within reach can swap the failed module on site.
This is why the service contract, not the headline spec, is where uptime is actually won or lost. A modest machine backed by a same-week parts and technician network will outperform a faster one backed by nothing local.
EDITORIAL RULE
Sourced or silent. Every figure on this page carries an attributed, dated source. Where we could not source a number, we left it out rather than estimate it.
SOURCES
- 01Manufacturer service SLAs2026
- 02Site commissioning logs2025–2026
- 03Operator interviewsMilan, London, Abu Dhabi · Q1 2026
- 04N.O.N. Humanoid vetting notesFactory audits · 2026
A QUESTION
One person answers — from the first email to the room.
EMAIL
WECHAT
nh_alliance
LINKEDIN
