Factorio Gleba Guide: Spoilage Loops, Defense & Ag Science

Factorio Gleba Guide: Spoilage Loops, Pentapod Defense, and Agricultural Science

Last verified: June 12, 2026 for Factorio: Space Age 2.0. The announced 2.1 update changes space logistics, not Gleba mechanics.

Gleba is the planet that breaks Nauvis brains. Every habit the base game teaches (buffer everything, stockpile intermediates, build big and let it sit) is precisely wrong on a planet where the products rot. The players who hate Gleba are almost always fighting the spoilage system; the players who love it stopped fighting and redesigned. This guide is the redesign: how the timers actually work, the one base-layout rule that matters, what the pentapods respond to, and how to ship agricultural science off-world before it composts itself.

Spoilage: the actual rules

• Most biological items carry a spoil timer, from a few minutes to roughly 2 hours depending on the item. When it expires, the item becomes spoilage (which is itself a resource: nothing is wasted, just demoted).
• Nutrients spoil in 5 minutes. Nutrients are also what every biochamber eats to run. This pair of facts is the whole planet: your machines run on fuel with a 5-minute shelf life.
• Freshness propagates. Products inherit spoilage state from ingredients, so a line fed with old fruit produces short-lived outputs. Speed is quality on Gleba.

Every spoil timer in Space Age, in order

Thirteen items in the Space Age data set carry a spoil timer, the same thirteen with the same timers in 2.1.23 as in 2.0.77. This is all of them, shortest first, with what each one turns into and its stack size:

Three things fall out of that table that change how you build. The two bacteria at one minute are the tightest deadline on the planet, and they spoil upward into ore rather than into spoilage, so a bacteria line that stalls loses its multiplication step but not its material. Mash at 3 minutes is shorter than jelly at 4, so a bioflux line fed from a shared belt starves on the mash side first. And the agricultural science pack itself has a 60 minute clock that starts the moment it is crafted, which is the real constraint on the ship-it-off-world section below: a pack that waits 40 minutes for a rocket arrives with two thirds of its life gone.

The one base-design rule: keep everything moving

The community consensus after a year and a half of Space Age is near-unanimous and worth stating as a law: no chest buffers between agricultural towers and first-stage processing. The working shape is a flow design: towers adjacent to biochambers, short loop belts, products consumed at the rate they are produced, and deliberate overflow disposal (burn it or recycle it) instead of storage. A belt loop with a filtered spoilage exit is the Gleba equivalent of main bus design: the pattern everything else hangs off.

Two practical corollaries:

• Every biochamber cluster needs a local nutrient loop. A bioflux-fed biochamber wants nutrients at about 15/minute, and with the 5-minute spoil window, central nutrient distribution does not survive belt latency. Make nutrients where they are eaten.
• Build the restart path before you need it. A fully stalled nutrient loop is a deadlock: biochambers need nutrients to make nutrients. Keep the spoilage-to-nutrients bootstrap recipe available and a manual kick procedure documented; on multiplayer servers, "who knows how to restart Gleba" should have at least two answers.

Where the 15 nutrients a minute comes from

It is not a community estimate, it is division. The shipped biochamber draws 500 kW while it is working, and a nutrient is worth 2 MJ of fuel. Half a megajoule per second against 2 MJ an item means one nutrient is four seconds of running, or 15 a minute at full load, before any module changes the draw. Nutrients also sit in a fuel category of their own in the data set, and nothing else in the game carries that category, so there is no coal-in-the-furnace fallback here: if the nutrient belt dries up, the machine stops.

The biochamber's other numbers are worth having next to that: crafting speed 2, four module slots, 300 hit points, and 50% productivity built into the machine itself before you put a single module in it. That base productivity is the reason Gleba lines look wasteful and are not.

The nutrient recipes, ranked by what they actually yield

There are five ways to make nutrients and they are not close to equivalent. Read the yield column before you decide which one your loop runs on:

The bootstrap recipe is 80 times worse per input item than the bioflux one, and that is the whole point of it: spoilage is the one input a dead base still has. Run it to get a single biochamber breathing, then hand the loop back to bioflux. Note also what the game calls these recipe categories. Spoilage, fish and biter egg nutrients sit in a category named organic-or-assembling; the bioflux and yumako mash recipes sit in plain organic. The naming is the hint: the restart recipes are the ones that do not need the organic machine you are trying to restart.

For reference, the chain above it: 15 yumako mash plus 12 jelly makes 4 bioflux in 6 seconds; one yumako processes to 2 mash in 1 second; one jellynut processes to 4 jelly in 1 second; and both processing recipes return a seed 2% of the time, which is the replanting supply. Burning spoilage is also a real option, at 6 spoilage to 1 carbon in 12 seconds, and spoilage itself carries 250 kJ of chemical fuel value if you would rather burn it than compost it.

Pentapods: defense scales with farming

Gleba's enemies respond to spores, emitted by agricultural towers when they harvest, drifting and accumulating exactly like pollution on Nauvis. The loop is mechanical: more farms → more spores → bigger, more frequent pentapod attacks. Three planning consequences:

• Defense is a function of production. Size your walls and turret lines for the farm you are about to build, not the one you have. Every agriculture expansion is also a military expansion.
• Stompers change the wall math. The large pentapods step over naive single-wall designs; depth (layered walls, spaced turrets, repair coverage) beats thickness.
• Eggs are an inside threat. Pentapod eggs, required for biochambers and agricultural science, hatch if they spoil, inside your base, next to your production. Egg-handling areas get their own turrets and repair packs, pointed inward. Treat eggs like live ammunition on a conveyor: keep the flow fast, never let them accumulate, and assume the occasional hatch.

The egg line, in numbers

The egg recipe is the most quietly alarming thing in the data set. One pentapod egg plus 30 nutrients plus 60 water makes two pentapod eggs, in a 15 second craft. The egg's spoil timer is 15 minutes. So the line doubles roughly every fifteen seconds and every unit on it is on a fifteen minute fuse, and it is not optional: an agricultural science pack is one bioflux plus one pentapod egg in a 4 second craft, so the pack rate is the egg rate, one for one. A biochamber costs an egg to build too, along with 5 nutrients, 20 iron plates, 5 electronic circuits and a landfill.

The practical reading: size the egg buffer in seconds, not in stacks. The egg stack size is 20, the smallest of any spoiling item on the planet, which is the game quietly telling you the same thing. If a downstream stall parks a chest of eggs for a quarter of an hour, that chest is the attack.

What is actually coming at you

Hit points from the Space Age 2.1.23 data set (unchanged from 2.0.77 for these five), so you can size a turret line against real numbers rather than vibes:

The premature variants that come out of a hatched egg carry the same hit points as their grown counterparts, so an accident inside the walls is not a softer fight than one outside them. For comparison, a big biter on Nauvis is 375 and a behemoth is 3,000, so individual wrigglers are not the problem. Volume driven by your own spore output is.

On the turret side, the durability ladder in the same data set runs gun turret 400, laser turret 1,000, flamethrower turret 1,400, rocket turret 1,500, tesla turret 2,000, artillery turret 2,000, railgun turret 4,000. 2.1 raised two of those: the rocket turret was 400 and the tesla turret 1,000 in 2.0.77. A gun turret is 10 iron gear wheels, 10 copper plates and 20 iron plates, which is why the honest early Gleba answer is more gun turrets rather than better ones. The railgun turret, at the other end, wants 100 quantum processors, 50 superconductors, 30 tungsten plates, 20 carbon fiber and 100 fluoroketone, which is four planets' worth of supply chain before you place one.

Agricultural science: produce-and-ship

Agricultural science packs spoil like everything else, and they gate real endgame research (asteroid productivity, railgun damage among them). The pattern that works: continuous production at whatever rate your platform logistics can lift, prompt launches, and an accepted spoilage percentage as a line item rather than a failure. Chasing 100% freshness costs more in design complexity than the lost packs are worth: ship fast, ship often, and let the loop's excess feed the nutrient bootstrap.

How much does the planet owe the rest of the run? 34 Space Age technologies list agricultural science among their ingredients, and the finite ones add up to 42,950 packs. The biggest single line items are epic quality and legendary quality at 5,000 units each, then captive biter spawner, Aquilo discovery and the second belt capacity upgrade at 3,000, spidertron at 2,500, and a block of 2,000-unit technologies: railgun, fusion reactor, foundation, overgrowth soil, advanced asteroid processing and the first belt capacity upgrade. In 2.1 promethium science dropped from 2,000 to 1,000 units and a new 1,000-unit technology for reaching the solar system edge joined the list, which is why there are 34 where 2.0.77 had 33 and the total did not move. Belt stacking and the quality ladder both being on that list is the part people underestimate: Gleba is not a side planet you visit for a badge, it is a prerequisite for two of the biggest quality-of-life upgrades in the expansion.

One pack takes one pentapod egg, so that total is also an egg budget: up to 42,950 of them, before the biochamber's built-in 50% productivity takes a bite out of the bill. Worth remembering the next time the egg line feels like an overbuild.

Multiplayer and server notes

• Gleba is the best "second planet owner" assignment. Its systems are self-contained, and one player who internalizes the no-buffer rule outperforms three tourists. The division-of-labor logic from our science progression guide applies doubled.
• Spoilage ticks are real computation. Heavily-belted Gleba megafarms add UPS load from item-state updates; if your server is already tight, see platform optimization and hardware sizing.
• There is a dial if your group hates it. map-settings.json ships a difficulty_settings object containing spoil_time_modifier, default 1. It scales spoil timers, so a group that wants Gleba's design puzzle without its stopwatch can turn it up rather than skipping the planet. See server settings for where the file lives.
• Save before the first egg line. On servers with autosave intervals tuned long for performance, shorten them while the team learns egg handling. The first inside-the-walls hatch is a rite of passage; losing an hour to it is optional.

Taking a group to Space Age? Host a Factorio server with Supercraft: the UPS headroom for the planet where nothing is allowed to stand still.

Looking for managed Factorio server hosting? Supercraft runs Factorio dedicated servers with daily backups, instant setup, and 4 region options.

Frequently asked questions