Factory Layout Principles in Satisfactory

Most factories get rebuilt not because they were badly designed but because they could not be extended. The difference between the two is a handful of decisions made before the first machine goes down.

The three mistakes that cause rebuilds

Building tight. A perfectly packed production block works exactly until you need one more machine. Leaving a machine’s width of empty space at the end of every row costs nothing and is the difference between adding and rebuilding.

No belt corridors. If belts were routed around machines as you placed them, adding a line later means threading it through a maze. Deciding where belts run before placing machines is the single highest-return habit in factory building.

Inconsistent heights. Choosing a floor height per project means every junction between two areas needs a ramp or a lift. Pick one height for factory floors and one for long-distance runs, and use them everywhere.

None of these are aesthetic concerns. They are the reasons a factory becomes unworkable.

Design in modules, not in factories

The unit worth thinking in is one recipe’s worth of machines with everything they need: input belt, output belt, power, lighting.

Get that unit right once, then repeat it. Three advantages follow:

It fits terrain. A small module can be placed where the ground allows; a single large layout cannot.

It is blueprintable. The Blueprint Designer Mk.1 takes a 32 metre cube, which is roughly a four-by-four foundation area with machines on it. Designing to that size from the start means your modules are stampable when you unlock it at Tier 4.

It is diagnosable. When output drops, the problem is inside one module rather than somewhere in a hundred-metre hall.

Manifold or balanced, decided once

For feeding a row of machines, a manifold — one belt past the row with a splitter at each machine — is almost always right. It reaches the same steady state as a balanced tree, costs a fraction of the space and materials, and survives you adding a machine later.

The exception is a deliberately undersupplied line, where a manifold starves the far end permanently while a balancer shares the shortage evenly.

Decide this once as a house rule rather than per row. Mixed approaches in one factory are harder to reason about than either approach used consistently.

Sizing from the constraint, not from the machine

Work backwards. The question is not “how many Constructors fit here” but “what does the thing I am feeding actually need”.

A worked example. Your smelters need 90 iron ore per minute. A Miner Mk.2 on a normal node gives 120/min. Rather than running the miner flat out and backing up 30/min, underclock it to 75% — power scales by an exponent, so the saving is disproportionately large, and the line runs steadily instead of surging.

The same logic applies at every stage. Matching supply to demand beats maximising supply almost everywhere, because surplus with nowhere to go stalls the producer anyway.

Vertical or horizontal

Vertical factories look better and use less ground. They are also harder to debug, because you cannot see a whole line at once.

The compromise most good builds settle on: production floors stacked, each floor readable on its own. Conveyor Lifts move items between floors in one tile’s footprint at the same throughput as belts, so verticality costs you nothing mechanically.

What it does cost is access. If you cannot walk beside a belt, you cannot trace it, so leave walkways on any line you expect to maintain.

Enclosed or open

A tension worth resolving deliberately rather than discovering later.

Fully enclosed factories look excellent and are miserable to debug — every belt is behind a wall. Fully open ones are easy to work on and look like scaffolding.

Enclose the outside, leave the inside open. Walls and a roof for the silhouette, interior floors visible so a line can be followed from input to output. Access panels at the points you know you will revisit cost nothing.

Plan for the next tier, not this one

Every phase asks for more of what the previous one asked for. A line built to exactly satisfy the current requirement will be undersized for the next.

Planning ahead does not mean building double capacity. It means leaving room: extending the foundation platform past the last machine, running the belt with spare throughput, and putting the manifold on the side you can add to. None of that costs meaningful materials.

What to do with a factory that already sprawls

Most people ask this fifty hours in, and the answer is usually not to rebuild.

Build the next site properly and leave the old one running. Early factories get superseded anyway once you have better recipes and higher belt marks, and retrofitting a sprawl costs more than building fresh.

If you do want to improve an existing site cheaply: add foundations underneath to create a floor, add lighting, and pick one colour to unify the machines. Those three changes cost very little and do most of the work.

Where to put things relative to each other

Two placement rules that resolve most layout questions.

Process at the source, assemble at the convergence. Smelting belongs next to the miner, because ingots are fewer items than ore and plates are fewer still. Assembly belongs where its several inputs meet, because moving all of them to several places is worse than moving them to one.

That single principle explains most well-organised factories: extraction and first-stage processing scattered across the map at the nodes, manufacturing concentrated in one place.

Put storage where you build, not where you produce. Concrete, foundations and structural parts are consumed wherever you happen to be constructing. A storage wall next to your main build area saves more walking than one next to the smelters that made it.

Labelling is not optional at scale

Beyond a certain size you stop remembering what a belt carries, and the recovery cost is high — tracing a line by walking it is most of factory debugging.

Three cheap habits:

Signs on every line entering or leaving an area. They cost almost nothing and they work from a distance, which colour does not.

Colour-code pipes by contents, using the game’s own colour values so the code means the same thing everywhere.

Name your stations and containers. Typing is tedious, especially on a controller, and it is still faster than opening six containers to find the right one.

Common mistakes

Placing machines before deciding where belts run. Everything after is a compromise.

Packing a block with no room at the ends. One more machine means a rebuild.

Choosing a new floor height per project. Every junction then needs a ramp.

Maximising extraction instead of matching demand. Surplus stalls the producer and costs power.

Enclosing a factory you will need to debug. Outside walls, open interior.