I never thought we could afford a tending robot. The spindle hours said otherwise

October 1, 2026 · 4 min read

A job shop owner looks at a machine tending cell and sees a robot replacing an operator who costs less than the robot. That is the wrong frame. The return is unattended spindle hours sold at the shop rate. Here is a worked example that starts from that number.

The short answer

In the example below, a tending cell that adds 20 unattended hours a week on a machine with real backlog pays back in about two and a half years at the midpoint of the planning range, and in about a year and a half if the cell can run 30 unattended hours. The same cell on a machine without backlog has no payback, because there are no extra hours to sell.

The shop, as a composite example

This is a worked example with made-up but realistic inputs, not a client story and not a quote. Every figure on this site is an illustrative planning range subject to engineering verification. Put your own numbers in the same slots and the arithmetic is yours.

Picture a 22 person shop with a busy three axis mill that cuts about 30 hours a week with an operator standing by, and a stack of repeat work that keeps slipping. The shop cannot hire a second shift. The owner assumes a robot cell costs more than a machinist and dismisses it.

The inputs we used

Shop rate: 95 dollars an hour billed for spindle time. Contribution: not all of that is profit. After material, tooling and the share of overhead that scales with hours, we kept 60 percent, so each extra spindle hour is worth 57 dollars. Hours added: 20 unattended hours a week in the base case, running into the evening after the operator loads a table of blanks, for 48 working weeks, or 960 hours a year. High case: 30 hours a week, 1,440 hours. Cell cost: the machine tending guide here carries an illustrative equipment range of 50,000 to 110,000 dollars and integration of 20,000 to 50,000 dollars, including the automatic door and the interface. We used a 115,000 dollar midpoint. Operating: 6,000 dollars a year for grippers, maintenance and software.

How the arithmetic played out

Base case: 960 hours times 57 dollars is 54,720 dollars a year, less 6,000 dollars operating is 48,720 dollars. A 115,000 dollar project divided by that is about 2.4 years. High case: 1,440 hours times 57 is 82,080, less operating is 76,080, and the payback falls to about 1.5 years. If the project lands at the top of the ranges, 160,000 dollars, the base case is 3.3 years and the high case 2.1 years.

Now the version the owner was picturing. Replace the operator's hours during the day without adding spindle hours and the savings are whatever fraction of one wage the cell frees, perhaps 30,000 dollars a year, against the same 115,000 dollar project. Nearly four years. That is the arithmetic that convinces people robots are for bigger shops, and it is the wrong arithmetic.

What moves the answer most

Whether there is work to run in the unattended hours. Then how many unattended hours the parts and the table capacity allow before someone has to reload. Then the shop rate and contribution margin. The equipment price moves the payback by months. The backlog moves it from a project to a non-project.

What would have killed it

Parts that cannot be presented on a grid, so the table holds an hour of work rather than a shift. A machine without a practical automatic door. Operations between cycles the robot cannot do, such as deburring or inspection on every part. And the common mistake, choosing the machine with the easiest parts rather than the one with the backlog.

Run your own numbers

For each machine, estimate the hours a week it would cut if it never waited for a person, and whether you have the work to fill them. Multiply by your rate and your contribution margin. The machine with the biggest number is the candidate. Describe that machine and its parts to Joe, the concierge on this site, and it will put the figures in this frame, ask about the door and the workholding, and tell you if the table capacity caps the unattended hours.

Related on this site

A tending cell on a machine with a backlog pays back on hours you can sell. The same cell on a machine that sits idle part of the week pays back never.

Put it to the test on your own job

Tell Joe what you are running and he will apply this to your numbers, including telling you when a robot is the wrong answer.

More notes