Circuit Boards and Computers in Satisfactory

Computers are where the cost of a part stops being intuitive. One computer at the base recipe costs 72 units of raw ore, which is more than most players would guess by an order of magnitude, and the recipe you pick changes that figure by a third.

Circuit boards first

Recipe	Output	Ore per unit	Inputs
Circuit Board	7.5/min	10.0	15 Copper Sheet + 30 Plastic
Alternate: Caterium Circuit Board	8.8/min	4.71	12 Plastic + 38 Quickwire
Alternate: Silicon Circuit Board	12.5/min	5.72	28 Copper Sheet + 28 Silica
Alternate: Electrode Circuit Board	5.0/min	not resolvable	20 Rubber + 40 Petroleum Coke

All in an Assembler at 15 MW, so power is not a differentiator here.

Caterium Circuit Board halves the ore cost — 4.71 against 10.0 — and produces slightly more per machine. If you have caterium, this is close to a free upgrade, and it is the single highest-leverage alternate in the electronics chain.

Silicon Circuit Board is the throughput pick at 12.5/min, at a modest ore cost against Caterium. It suits a factory already producing silica for aluminium.

Electrode Circuit Board has no ore figure for the same reason Coated Cable does not: Petroleum Coke exists only as a byproduct. If your oil setup produces coke you need to consume, this converts it into circuit boards.

Computers

Recipe	Output	Ore per unit	Power	Inputs
Computer	2.5/min	72.0	55 MW	10 Circuit Board + 20 Cable + 40 Plastic
Alternate: Crystal Computer	3.3/min	52.0	15 MW	5 Circuit Board + 2 Crystal Oscillator
Alternate: Caterium Computer	3.8/min	84.4	55 MW	15 Circuit Board + 52 Quickwire + 22 Rubber

Crystal Computer wins on every measure at once, which is rare. More output per machine, 28% less raw ore, and it draws 15 MW against 55 MW — less than a third the power.

The catch is Crystal Oscillators, which are themselves a slow and expensive chain. If your oscillator line is already built, Crystal Computer is unambiguously the recipe to use. If it is not, you are choosing between building one chain to improve another.

Caterium Computer is the throughput pick at 3.8/min but costs more ore than the base recipe — 84.4 against 72.0. That is worth stating plainly because “alternate recipe” does not automatically mean “better”. It is the right choice only when quickwire is abundant and ore is not the constraint.

The compounding problem

Ore per unit compounds through this chain in a way that is easy to underestimate.

A computer at the base recipe needs 10 circuit boards. Each of those costs 10 ore. That is 100 ore of circuit boards alone, before the cable and plastic — and the total lands at 72 because the figures already account for the full expansion.

Switch the circuit boards to the Caterium alternate and their contribution roughly halves. Switch the computer itself to Crystal and the whole thing drops again. The two changes together are worth far more than either alone, which is the general shape of optimising a deep chain: the savings multiply rather than add.

Sizing a computer line

Computers arrive as a phase requirement and the numbers are unforgiving.

At the base recipe, 2.5 computers per minute needs one Manufacturer at 55 MW, fed by four Assemblers making circuit boards, fed in turn by copper sheet and plastic lines. Ten computers per minute is four Manufacturers and sixteen Assemblers, before the copper and oil chains behind them.

Work backwards from the phase requirement and check the raw ore total against your actual node supply. This is the chain where a factory most often turns out to be node-limited rather than machine-limited, and finding that out after building sixteen Assemblers is a poor use of an evening.

Practical recommendation

If you have caterium: Caterium Circuit Board immediately. Largest single gain in the chain.

If you have an oscillator line: Crystal Computer. Better on ore, throughput and power simultaneously.

If you have neither: base recipes are fine. Build the caterium outpost before optimising computers, because circuit boards are upstream and the savings flow through.

If you run oil with surplus coke: Electrode Circuit Board turns a disposal problem into production.

Where computers actually go

Understanding the demand shape helps you size the line once rather than three times.

Space Elevator phases want computers in bulk, and the requirement steps up sharply between phases rather than growing smoothly.

Buildings. Hoverpacks, Blueprint Designers, Nuclear Power Plants and most large machines list computers in their build cost. This demand is bursty — nothing for hours, then eight computers for one building.

Supercomputers and the tier above, which consume computers as an ingredient rather than as a build cost, and that demand is continuous.

The practical consequence: build the line to satisfy continuous demand, and keep a storage buffer for the bursts. A line sized for peak demand idles most of the time; a line sized for average demand with a container in front of it covers both.

Plastic is the hidden constraint

Look at the base recipes again and notice how much plastic runs through this chain: 30 plastic per circuit board batch, 40 per computer batch. Plastic comes from oil, and oil setups are frequently the least developed part of a mid-game factory.

More than once, a computer line that will not scale turns out to be an oil problem rather than an electronics problem. Before adding Assemblers, check whether the plastic supply can feed the ones you have.

The Caterium Circuit Board alternate helps here too, since it uses 12 plastic where the base recipe uses 30 — another case of the benefit landing in a different chain from the one you changed.

Common mistakes

Assuming every alternate is an improvement. Caterium Computer costs more ore than the base recipe.

Optimising computers before circuit boards. Boards are upstream; fixing them improves everything above.

Building the line before checking node supply. 72 ore per computer adds up fast enough to expose a node shortage you did not know you had.

Ignoring the power column. Crystal Computer draws 15 MW where the others draw 55. On a row of Manufacturers that difference is substantial.

Reading a missing ore figure as zero. It means the input comes from a byproduct with no mining path.