Crane Time Is the Scarcest Resource: Which Construction Robots Need No Pick

Every trade on a site brings its own tools. The concrete crew has its formwork, the ironworkers their tying tools, the MEP trades their drills. Nobody queues for any of that. There is exactly one resource everybody shares and nobody can multiply: the hook.

A second tower crane is not a procurement decision. It is a decision about site setup, foundations, slew envelopes and often about the permit. Within a running project crane time is therefore fixed. The only thing that moves is the order — and in practice the day's pick list is the real programme, whatever the bar chart says.

That has a consequence for anyone introducing robotics here. The binding constraint is not what the robot can do. It is whether deploying it lengthens or shortens the pick list. A machine that performs a task superbly and needs the hook three times a day is competing with the structural works for the scarcest resource on the site — and it loses that competition reliably, because the structural works set the tempo.

There are two reasons, and only one of them is technical. The more important one sits in the German rulebook.

DGUV Vorschrift 52 “Krane”, § 29 governs whom an employer may put in independent charge of a crane at all: only insured persons “who have completed their 18th year of life”, “who are physically and mentally suitable”, “who are instructed in operating or maintaining the crane and have demonstrated their capability to him”, and “of whom it can be expected that they will reliably fulfil the tasks assigned to them”. It continues: “The employer must appoint crane operators and maintenance personnel to their tasks. For mobile power-driven cranes the employer must appoint the crane operator in writing.” Paragraph 2 exempts hand-operated cranes alone.

This is not a prohibition on automation. The rule does not say a crane may not move automatically; it says who may operate it. But it is written throughout around a named natural person — with an age, a suitability finding, a demonstration of capability and a signature. A machine cannot be appointed, and no parallel provision transfers that role to a system. Anyone planning an autonomous crane would first have to answer who exactly is being appointed in writing.

The second reason is the conformity route. The EU Machinery Regulation 2023/1230 applies from 20 January 2027 and already sits in the standards set on the werob construction page. It governs how machinery — lifting equipment included — is placed on the market. It does not create a route that makes the crane operator dispensable.

Taken together this is an organisational and legal statement about this planning cycle, not a technical forecast. It reads simply: whoever plans a site for 2027 plans with a crane operator.

Once the crane is accepted as a fixed quantity, the selection question gets precise. It is no longer “which robot needs no crane” — a yes/no question that throws away the interesting cases. It is: how many picks does this machine consume, and across how much output are they spread?

That sorts candidates into three classes, and the middle one matters most:

• No picks at all. The machine drives or walks itself to where it works and away again. It never appears on the pick list.
• One pick to mobilise, none in operation. The machine has to be lifted onto the deck once and then works a whole pour without seeing the hook again. One pick spread across thousands of work cycles is arithmetically close to zero.
• Competes continuously. The machine wants the hook every cycle, or it occupies the airspace the crane slews through.

The third class is not a disqualification. It only means the machine does not belong in the robot schedule but in the crane schedule, weighed there against the structural works like any other major-plant question. Skip that distinction and you have planned two long booms into the same air, and you find out on the day.

Layout. Dusty Robotics describes the FieldPrinter as a system that autonomously prints the layout from the BIM model onto the surface — concrete, subfloor, formwork, asphalt — at a stated accuracy of one sixteenth of an inch. HP correspondingly calls SitePrint an autonomous robot that prints site layouts directly onto the floor; the constraint there, on HP's own description, is an unobstructed line of sight to the total station. Both drive themselves. That is a sightline condition, not a crane question. What it means for trade coordination is set out in the piece on layout robots on site.

Rebar. The TyBot from Advanced Construction Robotics is the most instructive case, because it falls honestly into the middle class. ACR states that the robot works on a gantry adjustable from nine to 117 feet in span, running on the deck's existing screed rails or on rails ACR supplies, and that it identifies intersections by itself and ties at over 1,100 ties per hour. That gantry has to get onto the deck — and that is a pick. Exactly one. After it, the whole pour runs without another. Why the work itself then stops competing for crane picks and hoist access is set out in the piece on rebar-tying robots and is not repeated here.

Overhead drilling. Hilti markets the Jaibot under its own product name, “Semi-Autonomous Drilling Robot”. The worker positions the machine by remote control, and the holes within reach are then drilled automatically from the BIM data, referenced by the Hilti PLT 300 total station. So the Jaibot needs a person. It does not need a crane. Those are two different scarcities, and only one of them is contended.

Survey. Leica Geosystems describes the BLK ARC as an autonomous laser-scanning module for robotic carriers with which scan missions can be repeated autonomously; the first carrier is the Boston Dynamics Spot. The robot walks onto the deck and off it again. The time-series logic behind that is in the piece on autonomous ground survey with legged robots.

The value of these four cases is not primarily in hours saved. It is that they can be scheduled into the windows the crane does not want: nights, weekends, the hours around a pour. Anything that needs nothing from the hook is decoupled from the tempo of the structural works — and that is a programme advantage first and a labour advantage second.

Not every construction machine with a robotic element fits in this obligingly. FBR's Hadrian X is the counter-example, and for a reason that is easy to miss: it does not compete for the hook. It brings its own boom.

FBR states a 32-metre telescopic boom for the machine, with which walls up to three storeys high can be built from the roadside, along with a stabilisation technology that measures movement from wind, vibration and inertia and counteracts it in real time. A boom of that length works in exactly the volume a tower crane claims when it slews. So the conflict is not a timing conflict at the hook but a spatial conflict in the air — and that is the harder one, because it is exclusive for the entire duration of the deployment rather than for the duration of a pick.

The site preparation that follows from this is already described in the piece on bricklaying robots, down to mapping the exclusion zones against overhead lines, tower crane operating envelopes and adjacent scaffolds. The point here is a different and purely planning one: a machine with a 32-metre boom is a major-plant item in the programme, even if the brochure calls it a robot. It gets planned against the crane, not parked alongside the other robots.

Which brings us to the uncomfortable part. For the way up — material from the ground to the level where it gets installed — there is nothing autonomous you can procure today. Everything regularly named in this space is either an assist device a human guides, or a demonstration. Demonstrations do not go into a programme.

One exception can be named on the evidence, and it is exactly as narrow as it sounds. The LIFTBOT from Kewazo GmbH of Garching near Munich is a lifting device for material on scaffolding. Kewazo describes it on its own product page as “battery-powered and remote controlled”, and names deployments there at BASF in Ludwigshafen, at the Nobian site in the Netherlands, and a 72-metre scaffold together with Marr Scaffolding. The machine is real, it is deployed — and on the manufacturer's own description it is remote-controlled, not autonomous.

That is exactly how it should be named, and named that way it is useful: it takes material up a scaffold without a crane pick and without a human carrying chain. That makes it a hoist, not an autonomous crane. The distinction is already established across the werob desk — demolition robots have always been described there as remote-controlled rather than autonomous, and the LIFTBOT is held to the same standard.

For an operator this means the automation on a building site happens on the ground. What runs there is collected on the construction solution page — layout, rebar, drilling, survey, material routes on the level. Vertical transport is deliberately not in it, and that is not a gap in the offer but a statement about the market.

The decision actually being made here is a scheduling decision, not a technology decision. It has four parts.

First: the pick list before the robot list. Anyone who does not know how many picks a typical shift on this site consumes, and on what, is specifying robotics blind. That list already exists on every site, usually in the site manager's head.

Second: three numbers per machine. Picks to mobilise, picks in running operation, airspace required yes or no. The first selection round needs nothing more, and those three numbers sort candidates more reliably than any performance comparison.

Third: use the windows. Whatever runs without crane time belongs in the hours the crane does not want. That is the largest lever in the whole calculation, and it costs no additional hardware — only the willingness to stop aligning the robot schedule with the daily rhythm of the structural works.

Fourth: no autonomous crane in this plan. Plan one in and you are planning against § 29, and you will discover it in the appointment and approval process rather than on the site.

werob is manufacturer-independent and picks the platform from the task, not the other way round. Describe the site, the task and the crane loading in plain language: first spec in 48 hours, first robot on site in eight weeks.