The Five-Year Cost of a Service Robot, Line by Line

Why the Sticker Price Misleads

Almost every robotics comparison we are asked to review starts with unit price, and almost every one of them reaches the wrong conclusion. Hardware is typically 50% to 65% of what a service robot costs a facility over five years. The rest arrives afterwards, in categories that rarely appear on the quote that won the evaluation.

This is not vendors behaving badly. It is a genuine structural problem with how robots are sold: the purchase decision is made against a document that describes the machine, while the actual cost is driven by how the machine is operated, serviced, and eventually replaced. The purpose of this guide is to give you the full list so that the comparison is fair.

Capital and Landed Cost

Start with the obvious and get it complete. The line items that belong here are the unit price, any options or accessories, spare batteries, charging docks, inbound freight, customs duty and tariffs, and delivery to the floor rather than the loading bay.

Duty deserves particular attention right now. Tariff treatment on imported robotics has moved considerably, and a quote that shows freight but omits duty can understate landed cost by a meaningful margin. Ask whether the price is delivered duty paid. If it is not, ask what the duty rate is and who is the importer of record.

Service, Parts, and Wear Items

A service robot is a vehicle. It has consumables, and the honest way to budget for it is the way a fleet manager budgets for vans.

• Batteries. The most predictable recurring cost. Lithium packs in daily service degrade on a known curve and are typically replaced once inside a five-year window, sometimes twice on a two-shift operation.
• Wheels, castors, and drive belts. Cheap individually, frequent in aggregate, and heavily dependent on floor surface. Polished concrete and carpet transitions wear components at very different rates.
• Brushes, squeegees, filters, and pads on cleaning platforms. These are genuine monthly consumables and are frequently omitted from budgets entirely.
• Sensors. Low failure rate, high replacement cost. Worth knowing which sensors are field-replaceable and which require a depot visit.
• Labour. Whether a technician comes to you, whether travel is billed, and what the response time commitment actually is.

The question that separates a real service programme from a warranty is simple: when a robot stops working on a Tuesday morning, what specifically happens, and by when? A twelve-month manufacturer warranty covering defects is not a service programme, and the two are routinely presented as though they were interchangeable.

The Costs Nobody Quotes

These are the categories that turn a good business case into a disappointing one, and they are almost always the customer’s to bear.

Site preparation. Threshold ramps, door automation, a charging location with adequate power, and occasionally a wireless access point or two. Usually modest, occasionally not.

Integration. Elevator controllers, door access, and any connection to a nurse call, work order, or laboratory information system. This is the most commonly underestimated line on the list, and it is often billed by a third party rather than the robot vendor.

Training and re-training. The initial session is usually included. What is not included is the fact that in a department with 40% annual turnover, you are training a materially different team eighteen months later.

Internal ownership. Someone has to be responsible for the robot. In practice this is a fraction of a supervisor’s time, every week, indefinitely. It is real cost even though it never appears on an invoice.

Software and connectivity. Fleet management licensing, cellular data if used, and any charge for integrations or API access.

Exit. What the machine is worth at the end, what it costs to remove, and whether you own data and configuration that transfer to a successor system.

A Worked Five-Year Comparison

Take a single transport robot in a mid-sized facility, running one shift. Purchased outright, a representative five-year profile looks like this. Landed hardware including duty and freight sits at the top. Add site preparation and integration in year one. Add an annual service agreement, typically 10% to 18% of hardware value depending on coverage. Add one battery replacement around year three. Add consumables at a modest monthly rate. Add a fraction of an internal owner’s time throughout.

The pattern that emerges is consistent: year one is roughly 60% of the five-year total, and the remaining four years carry a steady annual load of service, consumables, and attention. Facilities that budget only for year one are surprised in year two, every time.

Under a Robots-as-a-Service structure the shape changes rather than the total necessarily falling. Capital moves to operating expense, the peak in year one flattens, and the obligation to keep a working machine on site transfers to the provider. Whether that is cheaper depends on your cost of capital and your appetite for owning a fast-moving asset. Whether it is more predictable is not really in question.

Reading Two Offers Side by Side

Normalise before you compare. Put both offers on the same five-year horizon, with the same shift pattern and the same assumed utilisation. Move every cost into the same six buckets: capital and landed, service, consumables, integration, internal effort, and exit.

Then ask three questions of each. What is included by default and what is quoted per incident? What happens to the price at renewal, and is there a cap? And who carries the risk if the platform is discontinued or cannot be replaced like for like?

That last question has become considerably more important in the past year, and any vendor who cannot answer it clearly has given you useful information about the offer.