Power ceiling
How much fleet PR each race can sustain indefinitely. Land comes from what clustering can actually reach, income from that land at the market prices you pick, and upkeep from live ship stats — sustainable PR is income divided by upkeep per PR.
Market and fleet
Speed is velocity — PR fielded per turn (pr × build rate), split at the roster median of 40,800 PR/turn. Typical and the two bands use each race's median upkeep per PR; best case uses its single most upkeep-efficient ship. Border buffer holds 6 single-planet colonies back as sacrificial buffer — you can lose 3 per attack, so it takes two hits before a cluster is exposed.
Viral research picks — default: 3 per donor race, most widely flown first
Viral researches ships from Terran, A.Miner and Marauder, so reading its own catalog gives three native ships and badly understates what it actually flies. These are the 29 researchable ships observed across every Viral empire on record, ranked by how many distinct empires field each. Viral's 3 native ships are always in the pool and aren't listed — they're owned, not chosen. Two rules are enforced: at most 3 ships per donor race (9 borrowed slots, natives free) — and picking a speed band assumes Viral commits all nine of those slots to that doctrine, so Slow and Fast are different repertoires, not two slices of one. Upkeep is billed at the Viral race modifier rather than the donor's — researching is a build, not a capture, so a Viral flying a Marauder ship pays Viral's rate, about 3.7× what a Marauder pays.
Currently driving the bands — slow: Tiger, V.Borrelly, B.Cronus, M.Epidote, M.Illite, T.Fenrir, M.Chlor, T.Garuda, Bullhead, Shark, T.Kalieum; fast: Nirvana, V.Triton, A.Aragonite, Angel, Tyr, T.Empereur, M.Lyth, Pyth, Thresher, M.Calcite. 20 ships excluded by the 3-per-donor rule.
Upkeep modifiers — what-if: rescale a race's ship upkeep and watch the fleet it can hold
Upkeep is charged at the owning race's modifier, so sustainable fleet is inversely proportional to it — halve the modifier and the race holds twice the PR on the same income. These are the live values; drag one to see what a rebalance would do. Income is unaffected, so nothing here touches the land or credits columns.
Sustainable fleet
at CG 5 · food 8 · minerals 1,200| Race | Planets | Net credits/turn | Income | Sustainable PR | |||
|---|---|---|---|---|---|---|---|
| Collective | 1,222 | 156,292,920 | Agriculture 2531/2 | 407M | |||
| Terran | 831 | 318,425,838 | Industry 2541/5 | 430M | |||
| A.Miner | 456 | 510,322,893 | Mining 2531/3 | 452M | |||
| Marauder | 456 | 55,794,192 | Agriculture 2481/2 | 300M | |||
| Viral | 390 | 89,241,600 | Tax 2551/3 | 102M | |||
| Guardian | 206 | 8,105,600 | Tax 255 | 76M | |||
Where the land comes from
clustering headroom, not the colony capForming a cluster consumes sub-colonies that must be held at the same time, so building one top-tier cluster needs a deep stack of scaffolding slots on top of everything already parked. That staging cost — not the colony cap — is what caps the fleet. Attacking for planets grants +3 over the cap, and that transient headroom is what makes the last cluster reachable.
| Race | Colony cap | Ceiling | Staging cost | Best holding |
|---|---|---|---|---|
| Collective | 13 | 16 | 13 | 4×256p + 3×64p + 6×border |
| Terran | 15 | 18 | 13 | 6×125p + 3×25p + 6×border |
| A.Miner | 12 | 15 | 13 | 3×125p + 3×25p + 6×border |
| Marauder | 12 | 15 | 13 | 3×125p + 3×25p + 6×border |
| Viral | 12 | 15 | 10 | 6×64p + 6×border |
| Guardian | 10 | 13 | 13 | 1×125p + 3×25p + 6×border |
Assumptions and what this model leaves out
Plunder is excluded. Marauder carries a +1,900% plunder modifier and Collective +1,100%, and neither is land income. Marauder's figure is the floor for a race built entirely around taking other people's credits, not an estimate of what it actually fields.
Land is fixed at 800 per planet. That one isn't a knob — the income figures were swept at it, so changing it would invalidate every stored number rather than rescaling them. Border posture is a knob because both settings were swept.
Each colony uses its own race's planet family, because the modifiers differ enormously: a standard Cluster Lvl 3 is agriculture +45% while Collective's Similare C.5 is −52% and Viral's Tainted C.4 is −35%. Captured clusters use the Infected family, the only one with positive agriculture and industry — which is what makes stolen land genuinely better than Viral's own. The six border colonies produce nothing at all: they're sacrificial buffer, and nobody develops a planet they expect to lose.
Research is 255 total across all five axes, not per axis, plus reallocatable mission grants (Terran +6, A.Miner +5, nobody else). At that budget every race specialises into essentially one income channel. Loyalty caps include mission rewards; Guardian starts at 2,500 and reaches 3,800, still short of everyone else's base.
Ships: Neutral ships, Starbases and Scouts are all excluded. Upkeep per PR is read live, so a rebalance moves these numbers without the income grid being rebuilt.
Viral is measured differently, on purpose. Every other race's bands are the median of its own catalog. Viral researches ships from other races, so its catalog is three native ships and reading it that way flatters Viral badly. Its bands instead come from the ships Viral empires have actually been observed fielding, natives included. Slow and Fast each assume Viral spends all nine research slots on that doctrine — three ships per donor race, most widely flown first — so they're separate repertoires rather than two halves of one. Typical is the mixed repertoire Virals are actually observed running. Best case is the cheapest ship in that legal mixed repertoire, the same definition as every other race.
Upkeep follows the owner, not the ship. Researching a donor ship is a build, not a capture, so a Viral pays the Viral race modifier on it. That matters enormously in both directions: Marauder's modifier is 1.9 against Viral's 7.0, so Marauder ships cost a Viral about 3.7× what they cost a Marauder, while A.Miner ships (modifier 10.1) come in around 0.7×. Reading the stored native upkeep would have made Viral look far cheaper than it is, on exactly the cheap Marauder ships it most wants.
Only single-planet NPC wins are counted — winning an existing cluster is possible and is excluded entirely. That understates Viral in particular, which can infect captured clusters and keep their original planet count, bypassing its own 64-planet ceiling.