Nothing in Satisfactory is a character build. What you are building is a factory, and the decision that shapes it more than any other is which alternate recipes you accepted from hard drives.
Alternate recipes are not upgrades. Some are worse. They trade one constraint for another, and knowing which trade you are making is the whole skill.
Alternate recipes are the real tech tree
Hard drives come from crashed pods and are researched in the MAM, giving you a choice of alternate recipes. Once taken, the others in that batch are gone until you scan more drives.
The categories worth understanding:
Ore-efficient recipes get more product from less raw material. Take these when your node capacity is the limit.
Power-efficient recipes use less electricity per unit. Take these when your grid is straining.
Simpler recipes skip an intermediate step entirely, which reduces your building count and your headaches, often at some cost in raw efficiency.
Early on, ore is not your constraint and building complexity is. Take the simpler recipes and stop optimising for material you have plenty of.
| Alternate | Why it is worth taking |
|---|---|
| Casted Screw | Screws straight from ingots, deleting the worst belt bottleneck in the early game |
| Steel Coated Plate | Reinforced plates without the screw chain entirely |
| Stitched Iron Plate | Cheaper reinforced plates, more wire |
| Solid Steel Ingot | Steel from iron ingots, far better ore efficiency |
| Coke Steel Ingot | More steel per iron ore if you have coal spare |
| Recycled Rubber and Plastic | Turns the oil byproduct problem into the solution |
| Diluted Packaged Fuel | Much more fuel per crude, at the cost of a fiddly loop |
Casted Screw alone justifies scanning hard drives. Screws are the single most tedious intermediate in the game and that recipe removes them from your life.
Ratios, and why yours are always slightly wrong
Every machine has a fixed input and output rate, and chaining them cleanly means matching those rates.
The clean approach is to calculate the ratio for a production line and build exactly that many machines. The practical approach is to overbuild the input side slightly and let a buffer absorb the difference, because perfectly balanced lines stall the moment anything upstream hiccups.
Overclocking with power shards raises a machine's rate at a disproportionate power cost. Underclocking does the reverse and is genuinely useful: two machines at 50% use less power than one at 100% for the same output.
Manifolds beat balancers for almost everything
You can split a belt evenly with a balancer, which is complicated, or run a manifold, which is one belt feeding a row of machines in sequence.
Manifolds take time to fill and then run at full rate indefinitely. They are simpler to build, easier to extend, and the only real downside is the startup delay.
Build manifolds. The balancer is a solution to a problem you probably do not have.
Power is a cliff, not a slope
Run out of power and the whole grid trips, including the machines producing the fuel for the power. Restarting a tripped nuclear or fuel grid by hand is genuinely unpleasant.
Build well ahead of demand, and keep a small isolated biomass or battery buffer that can restart the main grid if it goes down.
The progression is biomass, coal, fuel, then nuclear. Coal is the first stable plateau and the point where the game opens up, so get there without over-engineering the steps before it.
Build vertically and leave room
The most common regret is a first factory built flat and tight, with no space to extend when a later tier demands more inputs.
Leave gaps. Build up. Foundations are cheap and the alternative is demolishing a working line because you need one more constructor.
The Satisfactory beginner's guide covers the early milestones.
