Europe's First 3D Printed Office Expansion Project: How It Was Built in Austria

• 1
  Five things to know before you specify printed concrete
• 2
  What the Austria 3D printed office expansion project actually built
• 3
  Why concrete rheology decides whether the print succeeds
• 4
  Layer geometry, overhangs and what the printer cannot do
• 5
  Where CNC machining takes over from the printer
• 6
  Sourcing the metal side of a printed building
• 7
  When a printed concrete shell is the wrong choice
• 8
  How to plan a printed shell plus machined interface package
• 9
  Printed concrete vs cast concrete vs CNC machined parts

The project broke ground in Austria in 2022 and became the first 3D printed office expansion project in Europe. It added a two-story office wing to an existing commercial building. The structural shell came out of a gantry concrete printer running a cementitious mix through a 40–50 mm nozzle. Each pass laid a bead roughly 30–40 mm tall, and the machine stacked those beads into walls with a hollow core.

Three parties made it work. The construction group handled site logistics and scaffolding. The formwork supplier brought the concrete know-how. The printing technology partner supplied the gantry system and the control software. That split matters: no single vendor owned the whole problem, so the interfaces between them had to be drawn precisely before the first bead went down.

The finished building is not a novelty. It houses offices. That means it carries real loads, real fire requirements, and real thermal performance targets. Print speed was never the headline. The headline was that a printed shell passed European building review and got occupied.

For an engineer, the interesting part is the seam between printed concrete and everything else. The printer cannot hold a bolt pattern. It cannot cut a glazing rebate. It cannot hold a door frame square within 2 mm. Every one of those interfaces became a second operation, and that is where the real cost sits.

• 1
  StructureTwo-story office wing, printed load-bearing walls on a conventional slab.
• 2
  Printer classGantry system with a 40–50 mm nozzle, layer height 30–40 mm.
• 3
  Delivery modelConstruction group, formwork supplier and print technology partner working as three contracts.

Printed concrete has a narrow working band. It must pump easily, hold its shape the moment it leaves the nozzle, and still bond to the layer below. Those three demands pull against each other. A mix that flows well usually sags. A mix that stands up usually sets too fast to pump reliably over a long run.

The practical control is open time. At 20 °C a typical printable mix stays workable for 45–90 minutes. Print faster than that and you are fine. Stop for two hours and the material in the hose is scrap. On the Austria build, the schedule had to respect that window, which is why the print runs were planned around continuous shifts rather than convenient working hours.

Aggregate size sits at 4–8 mm for most building-scale printers. Larger aggregate raises compressive strength but blocks the hose. Fiber reinforcement at 0.5–1.5 % by volume is common because it lets the mix carry tensile stress across a cold joint without a full rebar cage in every wall.

The number that surprises people is shrinkage. Printed concrete dries faster than cast concrete because so much surface area is exposed. Restrained shrinkage cracks along layer lines if curing is not managed. Wet curing, coverings, or a curing compound applied within the first hours are not optional steps.

• 1
  Open time45–90 minutes at 20 °C; plan continuous pours, not stop-start shifts.
• 2
  Aggregate4–8 mm typical. Bigger aggregate raises strength but risks hose blockage.
• 3
  Fiber0.5–1.5 % by volume to bridge cold joints and control early cracking.
• 4
  ShrinkageHigher than cast concrete. Cure within hours, not days.

A concrete printer builds vertical walls well. It builds horizontal cantilevers badly. Without support, each layer must sit on the layer below with enough overlap to stay stable. In practice that caps unsupported overhang at roughly 10–20 degrees from vertical, depending on mix and layer height.

Curved walls are cheap. A gantry printer follows an arc as easily as a straight line, so the complex plan shapes that cost a fortune in formwork cost almost nothing in print time. That is the real economic argument for printed concrete: not speed, but formwork elimination.

Windows and doors are handled by leaving gaps and printing a lintel above, or by printing a temporary support and removing it later. The opening tolerance is where problems appear. A printed opening might be 20 mm out of position. The window frame supplier will not accept that. The fix is a cast or machined sub-frame that absorbs the deviation.

Internal voids are printed in deliberately. A wall printed as two parallel beads with a gap between them gives you a service cavity for conduit and insulation. It also reduces material use. The structural engineer has to confirm the wall still carries the load with that void present, which usually means a thicker overall section.

• 1
  Overhang limitAbout 10–20 degrees from vertical without temporary support.
• 2
  CurvesFree. Plan geometry is software, not carpentry.
• 3
  OpeningsExpect 20 mm position error. Use a machined sub-frame to absorb it.
• 4
  CavitiesPrint voids for services and insulation; confirm section thickness with the structural engineer.

Printed concrete gets you a shell. It does not get you a building. The gap between the two is filled by secondary operations, and most of them are metal.

Connection plates anchor the printed walls to the slab and to the roof structure. Those plates need flat datums and hole patterns held to ±0.1 mm, which is three orders of magnitude tighter than the print. We machine them from 6061-T6 or 304 stainless depending on the corrosion exposure. A printed wall might sit 10 mm off nominal; the plate is designed with slotted holes to swallow that deviation while keeping the bolted joint tight.

Window and door frames follow the same logic. The frame itself is extruded aluminum. The bracket that ties it to the printed wall is machined, often as a small family of parts in the same setup so the whole run shares one datum. On a two-story building you might need 40–120 of these brackets. That is well inside a single 5-axis run.

MEP penetrations are the third interface. Sleeves through a printed wall have to be positioned, sealed and fire-stopped. Machined collars with a machined gasket groove are more reliable than site-cut foam. They also install faster, which matters because the printed wall is already cured and unforgiving by the time they go in.

• 1
  Anchor plates6061-T6 or 304 stainless, ±0.1 mm hole position, slotted to absorb print deviation.
• 2
  Frame brackets40–120 pieces per building typical. One setup, one datum.
• 3
  Service collarsMachined gasket groove beats site-cut foam for fire-stop reliability.

If you are running a printed concrete project, the metal package is usually the schedule risk. Printed walls are on site early because the print is fast. The brackets, plates and collars that make the walls usable arrive later, and if they are wrong the building sits open.

The parts are small but the tolerance is tight. A connection plate with eight holes at ±0.1 mm is a normal 3-axis job. A bracket with features on four faces is a 5-axis job. Neither is difficult. What matters is that the shop understands the print tolerance it has to absorb, because the design intent is not a precision joint, it is a precision joint that stays precision when the substrate is 10 mm off.

Order quantity is usually awkward. One building might need 40 brackets. That is below most shops' comfort zone. We run from one prototype to 10,000+ part runs with no minimum order quantity, so a single building's worth of hardware is a normal order for us.

Materials for exterior connection hardware are usually 6061-T6, 304 or 316 stainless, or hot-dip galvanized steel. Finish matters more than alloy in most European climates. Anodizing or powder coating over a properly prepared surface will outlast a bare alloy choice every time.

• 1
  Schedule riskMetal package usually gates occupancy, not the print.
• 2
  Quantity40–120 brackets per building is normal; no minimum order quantity from us.
• 3
  Material6061-T6, 304, 316, or galvanized steel for exterior connections.
• 4
  FinishingAnodizing or powder coat over correct surface prep; alloy choice matters less.

Printed concrete is not a general replacement for conventional construction. It wins on freeform geometry and formwork elimination. It loses on anything that needs a tight, flat, repeatable surface.

If your building is a rectangular warehouse with a flat roof, cast and steel will beat the printer on cost every time. Printing shines when the plan has curves, when the wall count is high, or when formwork would be a large share of the budget.

If your program needs many identical units, the economics shift again. A printed shell still needs its metal interface package per unit, and that package does not get cheaper with repetition unless you tool it. At high volume, precast plus CNC-machined connections usually wins.

The honest boundary is this: print the shape, machine the function. When a design tries to make the printer do precision work, it fails at the interface and the cost lands on site labor. That is the lesson the Austria project encoded in its drawings, and it is the lesson worth copying.

• 1
  Choose printingCurved plans, high wall count, formwork-heavy geometry.
• 2
  Choose cast and steelRectangular, repetitive, flat-surface buildings.
• 3
  Choose hybridPrinted shell plus machined interface package, which is most real projects.