Sandustry — the Early Access walkthrough: Factory Tiers 1 to 7

Sandustry is a factory game where the ground itself is the resource. Nothing is an ore node on a map: every pixel of terrain is simulated, so digging out a hill gives you the sand it was made of, pouring water on that sand gives you wet sand, and shaking wet sand gives you gold. The factory is built out of that sand and rebuilds itself out of the by-products when it is done with them.

It matters that this is a falling-sand game rather than a belt-only one, because most of the difficulty in the first fifteen hours is not ratios — it is density. A line that works on paper stalls because the wrong material is heavier than the thing it needs to displace, or because a filter let a pixel through that then transformed inside it.

This page is the progression walkthrough: the seven Factory Tiers, what each one actually asks you to build, and the research bill for each. It is written for the Early Access build, which at the time of writing is 0.6.0 (21 October 2026).

What Sandustry is, and where Early Access is right now

The version history matters more than usual here, because several rules players learned at launch have since been replaced. The changes that alter how you build, and not just how the game feels:

Two of those are traps. If you are following an older guide, its gas behaviour is backwards and its pipe advice will produce a network that does not connect.

The rules the sand obeys

Almost every stall in Sandustry is one of five rules.

1. Density decides what sinks and what floats. Gold is the heaviest material in the game at 300 and always sinks; Water is 100; Lava is 200 as of 0.5.6. Two materials swap places by weight, not by which one you dropped first. This is why a shaker staircase works and a flat shaker field does not.

2. Matter type decides how it moves. Solids fall straight down. Slushy materials (Residue, Wet Sand, Wet Seed, Voidbloom) behave like wet sand and pile. Liquids spread and level. Gases rise and pool under ceilings — and as of 0.6.0 the lower the density the higher a gas rises, which is the inverse of how it shipped. "Wisp" materials (Amethelis, Dry Amethelis, Snow) drift sideways as they fall, so anything that catches petals needs a wide mouth, not a one-pixel funnel.

3. Three flags decide which machine cares. Burnable (Residue, Cinder, Water, Snow, Dry Amethelis, Vine, Moss, Florinol Soil) is what fire acts on. Wettable (Sand, Seed) is what water converts. Launchable is what a Launcher will throw — Gold and Residue are, Water and Lava are not.

4. Buildings filter by material, not by shape. A Shaker passes only Gold through itself and holds everything else on top. That single rule is the whole gold line: you do not sort the output, you let one material fall out of the bottom while the rest is pushed off the side.

5. Nothing in logistics needs power. Conveyor Belts, Launchers, Filters and Portals run continuously and consume nothing. Power is only for the buildings that refine, extract or store.

The Factory Tiers are the whole progression

Research is gated by Factory Tiers. A tier is a delivery target: you produce a quantity of work, the tier completes, and the next band of research opens. There are seven.

Two things to notice. First, every tier's task is the output of the technology you just unlocked, so the tier is a tutorial that pays you. Second, the targets are volume, not rate — a slow line that never stops beats a fast line that jams, and the fastest way to fail a tier is to build a beautiful factory that deadlocks once and is never noticed.

The quantities are exact from the game's own tier screen. Tier 7's target is not yet documented anywhere outside the game, and 0.6.0 has only been out a week at the time of writing.

Tier 1 — Wet Sand into Gold

Tier 1 wants 4,000 Wet Sand put through a Shaker. Everything else in the tier is logistics.

The chain, exactly as the game documents it:

1. Dig Dirt (4 HP) → Sand.
2. Water + Sand → Wet Sand. Wet Sand is denser than Sand (150 both, but Slushy against Solid), so

it does not float away, and it does not fall through water either.

1. Wet Sand on a Shaker → 1 Residue, plus a 25% chance of 1 Gold.
2. Only Gold passes through the Shaker. Residue stays on top and is pushed off the side extremely

slowly — the Shaker is effectively a very slow conveyor for anything it cannot pass.

1. Gold lands underneath, in your Collector. Spent gold is taken from Collectors automatically.

The one layout rule worth learning early: build Shakers in a descending diagonal, not a flat field. Wet Sand is denser than Residue, so on a staircase the incoming sludge pushes the residue off each Shaker and settles as a thin layer on the next one down. On a flat field the residue piles up until it covers the shaker and gold stops falling through.

The wiki's own overall arithmetic, for planning: roughly 2 Sand becomes 3 Gold once the full chain is running, with about 4 Residue needed per Burnt Residue later.

A note on the wet-sand ratio, because two official pages disagree: the Sand page states 1 Sand + 1 Water → 2 Wet Sand, while the Wet Sand page states a Sand-to-Wet-Sand ratio of about 4:1. Both cannot be right. The practical answer is that it does not matter much — water is recirculating and the bottleneck is always Shaker throughput — but do not build a water budget off either figure without measuring your own line.

Tier 1 research order

Tier 1 costs 800 Gold plus the unpriced Wall Light, which is deliverable inside the first hour.

Tier 2 — Burnt Residue and the 25-block drop

Tier 2 wants 3,000 Burnt Residue pressed. The step that catches people is that pressing is a fall, not a machine setting.

1. Residue + fire → Burnt Residue. Roughly 4 Residue per Burnt Residue. Fire can come from a

Flamethrower, Lava, a Pyro Dispenser or a Burner Belt — the Burner Belt is the automatable one, because it sets flammable pixels on fire while transporting them.

1. Drop Burnt Residue from at least 25 blocks of height onto a Kinetic Press → 1 Gold + 1 Seed.

A press under a short drop does nothing; the height is the recipe.

1. The Seed is the Tier 3 input, so this tier is where the loop from "sand" to "biology" starts.

This is the tier that most rewards a vertical build. You want a tower: residue burning at the top, 25+ blocks of clear shaft below it, Kinetic Press at the bottom, Collector under that. Launchers carry pixels upward and a tower with only the top launcher angled deposits out the side (hold Shift to force straight-up placement), so you can run a return line up the same shaft.

Tier 2 research is the expensive one relative to what you own — 5,550 Gold across eight nodes.

If the Flamethrower's "Water becomes Steam, returns as rain" line looks like a free water source, it nearly is: Steam rises, and with no ceiling it reaches the sky and condenses back into Water. Ceilings interrupt it, which is exactly what you want later when you need steam inside a machine.

Tier 3 — Flowers, Amethelis and 4,000 plants

Tier 3 wants 4,000 Flowers grown. This is where the game stops being about sand.

1. Seed + Water → Wet Seed. Wet Seed has the same density as Water (100), so it stays suspended —

it will not settle out of a water column on its own. Seed is 150 and does sink.

1. Wet Seed in a Planter Box → one Flower, guaranteed: 1 Wet Seed = exactly 1 Flower. A growing

plant absorbs extra Wet Seeds and grows one flower per seed, so overfeeding does not waste them.

1. A Flower is a Vine stem with a head of 4 Amethelis + 1 Gold. Break the head and it drops all

five. Since the tier counts flowers grown, not harvested, you can either run a continuous plant-and-harvest loop or plant a field and let it stand — the tier completes on growth.

Amethelis is a Wisp material: it drifts sideways as it falls. Anything you build to catch petals needs a wide intake or a fan, and the Aerokinetic Fan (Tier 3 research) exists for exactly this.

Tier 3 research totals 14,450 Gold, and it is the tier where the tech tree branches rather than stacking.

Tier 4 — Steam, Florin and Florinol

Tier 4 wants 10,000 Florin condensed into Florinol, and it is the first tier where the chain has a genuine thermal requirement at both ends: you need steam to dry, then fire to burn, then cold to condense. One Thermal Buffer feeds all three if you move it, and three of them is easier.

Two details worth knowing before you build it.

The Condenser pays gold. Two Gold per Florinol, on top of the tier progress. A Tier 4 line that runs continuously is also a gold line, which is how you afford Tier 5's 33,500.

Cold has two sources and one of them is a trap. A cold Thermal Buffer works; Snow dropped directly in works too, but Snow fed from inside a Thermal Buffer runs at 10% rate — no gain over using Snow directly. A Cryoblaster gives you infinite Snow, so Snow is usually the cheaper cold.

Tier 4 research totals 15,250 Gold.

Tier 5 — Peak Energy, 100,000 of it

Tier 5 wants 100,000 Energy. That is not a number you can hand-feed; it is a power plant.

There are three ways to store energy, and they are not equivalent:

The Power Brick is the upgrade that makes the tier finishable: 1,000 energy from one brick, and the copper comes back when you break it. Liquid Copper is 1 Copper → 1 Liquid Copper through a Smelter, which needs metal on top and heat from a Thermal Buffer on any other side (or Lava directly underneath). If you feed a Smelter from Lava rather than from heat in a Thermal Buffer, expect it to use about 2.5× as much Lava — the wiki's own figure, and the reason the buffer is the better investment.

Water and Steam are the infinite side. Water + fire → Steam; Steam under a ceiling pools where you want it; a Steam Turbine converts it at 10 energy per pixel and hands you Depleted Steam, which condenses back to Water. Close that loop and the tier completes on its own, which is what "Peak Energy" is asking for.

Tier 5 is the biggest research tier: 33,500 Gold across thirteen nodes.

Tier 6 — Aura

Tier 6 wants 10,000 Aura, and Aura is the only resource in the game you cannot dig, burn or condense. It comes out of a machine being fed by a machine.

Three rules make or break Tier 6:

• Auralite Crystal is immune to every player tool. Once the extractor crystallises it, you cannot

dig it, shoot it or melt it. The swarm is steered by clicking the extractor, not by mining.

• Wells must sit just outside the extractor's disc. Build them inside the radius and they do not

work; the extractor's reach is the boundary. Prismaline accumulates in storage while the extractor runs and is released by the Prismaline Well, so the well is part of the same footprint.

• The extractor can be removed mid-crystallisation (0.5.6). Crystals are buffered and re-released

as Auralite when you rebuild it, so a misplaced extractor is not a lost run — that was a real softlock before the patch.

Tier 6 research adds Synthesizer at 7,500 Gold, and the wiki's table stops there; 0.6.0 added two more nodes to the Aura tier along with the rest of Tier 7.

Tier 7 — sand refinement, and what 0.6.0 actually added

0.6.0 (21 October 2026) is the first patch that extends the progression instead of fixing it, and its own description of the new tier is a production chain for sand refinement "involving magnets, electrolysis, photosynthesis and more" — a deliberate return to the game's first material after five tiers of everything else.

These are the nodes past the Tier 6 gate, with the costs exactly as the 0.6.0 tree prints them:

That is 107,500 Gold across fourteen nodes. 0.6.0 shipped sixteen new nodes, so at least two are not in this list — and the gap is most likely the photosynthesis half of the chain, which none of the fourteen above covers.

The shape of the tier is three new mechanics and one bottleneck — Energy per pixel.

The tier is priced in Energy, not in belts. Both machines that do the new work charge 200 Energy per element — the Grinder for every element it grinds, the Electromagnet for holding its field. At that price a line running a few elements a second is drawing thousands of energy a second, so Tier 7 is gated on power throughput in a way no earlier tier was. Which is why 0.6.0 also ships a way to make power rather than capture it: the Electrolyzer turns Water into Hydrogen and Oxygen, and the Fuel Cell burns the Hydrogen back into Energy. The tier closes its own loop — electrolysis feeds the fuel cell, the fuel cell pays the Grinder and the Electromagnet. Build the power side first.

Gases become a real network. Hydrogen and Oxygen get their own pipe layer in Gas Pipes, which shares a tile with any other block and passes through everything except Artifact rooms. Because 0.6.0 also inverted the gas density rule, any gas column built before 21 October may now separate the other way round — that one change can make an existing gas loop behave backwards.

Magnetism becomes a material. The Electromagnet wants Magnetite fed in from above and 200 Energy, so it is the first building that asks you to route a solid and a utility into the same block. The Magnet is the hand version: a tool that pulls magnetic material to you instead of a building that moves it for you. Where Tier 5 taught you to move heat around, Tier 7 makes you move attraction.

Refinement turns back toward the ground. The Grinder works Quartz into Silica, which connects the snow and mountain biome 0.6.0 added to the production chain instead of leaving it as scenery. The Balancer is the first dedicated distribution building — and note its wording: it spreads elements below itself and adjacent balancers, so it belongs above the row it feeds, with its siblings side by side, not beside the line it is splitting. The Reaction Chamber is the multi-input reactor, filed under the Rheology group header, where elements are made to react inside the chamber to produce materials that no single machine in the earlier tiers could make.

And gold gets denser. Gold Granulation is a Synthesizer recipe that converts Liquid Gold into Gold Granulate. Density is the currency cap in this game — a Collector holds more gold the denser the gold is — so this is the node that makes the million-gold achievement practical rather than a wall of 62,500 Collectors.

The tree itself got easier to read in the same patch: it now has group headers drawn as yellow labels — Autobuild, Advanced, Research, Redstone, Rheology, Bombardier, Redeploy, Keyboard — and it zooms out far enough to see the whole thing at once. Plan the tier on the zoomed-out view before buying anything; the four nodes at 8,000 and the pair at 9,000 and 10,000 add up faster than the Tier 5 habit expects.

Still not documented anywhere, including here: what quantity Tier 7 asks you to deliver — the Viability panel in a completed save still lists the six pre-0.6.0 rows, so the gate may not be surfaced at all — and which of the 0.6.0 critters is progression-required; the Conservatory still reads 7/7, which is the pre-0.6.0 roster, so that creature is not tracked there either. The Reaction Chamber is the third unknown: its tooltip promises new materials without naming a single reaction, so its recipe set is in-game discovery. Two of 0.6.0's sixteen nodes are likewise uncatalogued. Everything above that is not in those four gaps was read off the live 0.6.0 build, not inferred.

Belts, launchers, filters and pipes: the rules that changed

Belts are horizontal. Direction comes from the direction you drag. There is no vertical belt: you use a tower of Launchers, which carry pixels up, or a Conveyor Portal pair. A launcher tower with only the top launcher angled deposits out the side; Shift forces straight-up placement. Mk.2 belts and launchers are documented only as "faster" / "more velocity" — no numeric speed is published anywhere, in-game or out, so plan with measurements, not with a number from a guide.

Filters act on material type, not on position. The basic Filter handles one solid type in Allow or Block mode; liquids and gases always pass. Two failure modes are worth knowing:

• A tall mixed pile can leave a lower-density allowed pixel stuck above denser ones, except within

about one block (4 pixels) of the Filter itself.

• A pixel that transforms while inside the filter becomes stuck and cannot be pushed out.

0.6.0 replaced the pile of ad-hoc interaction rules with one: two elements meeting across a filter boundary only interact if both elements and the output are allowed inside the filter. If a chain that used to react in a filter has gone quiet, that is why.

Pipes since 0.5.7 do not auto-connect. You must start or end a pipe run on an existing pipe tile, and building across a pipe creates a Pipe Bridge instead of a join. A fully submerged Pump delivers 16 pixels per tick, every Pump on a network feeds every Liquid Vent on it, multiple liquids in one network are distributed evenly, and a Vent prioritises the closest liquid. Lava cannot be pumped at all — it is "too hot to be transported".

Heat, cold and Lava

Hot Thermal structures only refine solid → liquid or liquid → gas; cold ones reverse it. That one sentence covers Smelter, Condenser and Snowmaker, and it is why a single Thermal Buffer can be a heater in one chain and a chiller in the next as long as you are not trying to do both at once.

Because Lava cannot be pumped, moving it is a three-step conversion: freeze to Scoria, dig to Cinder, re-melt with a direct flame. Thermal Relays move the heat itself across distances without moving any lava at all, which is almost always the cheaper answer.

Critters and the Conservatory

Critters are captured with the Corraller and recorded in the Conservatory, which pays tickets — and the tickets are where the game hides its logic gates. Piping and signalling is not in the tech tree; it is in the Conservatory.

You get one ticket per newly discovered critter type and more for extra captures, so the entire signal network is behind exploration rather than gold.

The Conservatory tops out at seven. In a completed 0.6.0 save it reads 7/7 discovered and 11/11 rewards unlocked, with 195 tickets left over and nothing to spend them on — and the seven are the original roster above. So 0.6.0's new critters are not tracked in the Conservatory: the ticket economy is finished the moment you have caught the seven, and the patch notes' claim that one of the new creatures is required for progression is not a Conservatory gate. Catch all seven for the signal network, then stop looking there for the new one.

The full research bill

Sums of the listed costs, from the wiki's table — which predates 0.6.0 and is marked work-in-progress, so treat Tier 6 and the Aura tier as floors rather than totals.

Tier 5 is 43% of the pre-0.6.0 bill, which is the shape of the game: the energy tier is the money tier, and everything after it is paid for by the factory you built there. Tier 7 then costs more than three times Tier 5 — 107,500 Gold for the fourteen nodes that can be read off the tree, and that is with two of the sixteen still uncounted. Do not arrive at it with an energy plant and no gold line.

What 1,000,000 Gold actually means

The Heliodyne Tycoon achievement asks you to hold 1,000,000 Gold in your collectors at once. The wiki's arithmetic on how big that is:

• Solid Gold: 62,500 full Collector blocks, which is a solid 250 × 250 wall of gold.
• Liquid Gold: 31,250 blocks — about 177 × 177, or 125 × 250 — because Liquid Gold is worth

double a Gold pixel.

That is the endgame of the gold economy, not a mid-game target. The step before it, Vaultkeeper, wants 100,000; Quartermaster wants 10,000; Gold in the Bin wants 1,000.

Achievements

Sixteen in total, and three of them are progression markers rather than optional:

How this guide was made, and what it does not claim

Every number here comes from three sources: the official Sandustry wiki (Hooded Horse's community wiki, all 198 pages, as it stood on 29 October 2026), the ten official Steam announcements from release on 13 August through 0.6.0 on 21 October, and — for the nodes the wiki has not caught up with — the live 0.6.0 tech tree, read node by node in a completed save and confirmed with the player who ran it. The walkthrough itself — the order of the tiers, the layout advice, which step is the bottleneck — comes from a full playthrough of everything in the Early Access build.

What that means in practice:

• Tier 7 is read off the live build, with four gaps left open. The fourteen node names, their

costs, and the Grinder, Electromagnet, Electrolyzer, Reaction Chamber and Balancer tooltips all come from the 0.6.0 build itself. What is still not visible anywhere — and is named as a gap rather than guessed at — is the tier's delivery quantity, which 0.6.0 critter gates progression, the Reaction Chamber's actual reactions, and two of the sixteen new nodes.

• The wiki's research table is marked work-in-progress by its own authors. The pre-0.6.0 figure is

a floor: Tier 6 is visibly incomplete and stops at a single node.

• The Sand-to-Wet-Sand ratio conflicts between two official pages (1:1 vs about 4:1). Both are

quoted above rather than one being picked.

• Gas densities are probably stale. 0.6.0 says it updated every gas density after inverting the

rule, and the wiki's values carry no revision marker, so treat the gas rows in the density table as pre-0.6.0 until measured.

Sandustry is in Early Access and its numbers move. When a patch changes a recipe, the in-game panels are the final authority and this page will be behind until it is rebuilt.