Strategy guides
Science Victory City Planning: Roles, the Production Core, and the Spaceport
A city-by-city plan for science victories in Gathering Storm: which cities carry research, which one carries the spaceport, era checkpoints to measure against, and what to do when the plan stalls.
What this guide assumes
- Standard game speed on a standard-size map. On faster or slower speeds every timing here shifts proportionally, but the order of decisions does not change.
- Difficulty in the Emperor-to-Immortal band. Below that, almost any plan works and this one simply works faster; on Deity, the same structure holds but you must buy military insurance earlier and accept later checkpoints.
- Gathering Storm rules with all DLC enabled, on the final patch. Power, geothermal fissures, and the laser-station finish all matter to this plan.
- Mostly peaceful geopolitics — no committed warmonger on your border. If you do have one, budget units before your third Campus, expect each era checkpoint to slip, and read the stalled-game section early rather than late.
One Empire, Two Races: Why City Roles Win
A science victory is two races run at once. Science per turn decides how early you unlock Rocketry and the projects behind it; production in one or two chosen cities decides how fast those projects finish. Empires that treat every city identically usually win the first race and lose the second: a dozen Campuses arrive on schedule, and then a Mars-length project exposes that no city was ever groomed to build anything heavy.
This is also why not every city needs a Campus. A Campus in a filler city with weak adjacency ties up production better spent on a Commercial Hub, a Harbor, or plain infrastructure, and its buildings add maintenance a small city struggles to justify. Past your best four to six sites, each additional Campus usually returns less than the same hammers invested in the cities that will carry your late game.
Assign roles deliberately instead. A handful of high-adjacency cities carry research. One production monster becomes the spaceport city. Trade-friendly cities fund the empire and feed domestic routes into the core. The rest exist to claim land, luxuries, and strategic resources — they want granaries and walls, not districts they will never leverage. Deciding these roles when you settle, not when you unlock Rocketry, is the cheapest improvement most players can make.
Placing Campuses the Map Actually Pays For
Campus quality is fixed the moment you place the tile, so spend real thought there. In Gathering Storm the heavy adjacency sources are mountain clusters, reefs, and geothermal fissures, plus bonuses from neighboring districts — a Government Plaza corner quietly upgrades an ordinary spot. Before committing, test candidate tiles in the adjacency calculator on this site rather than eyeballing it in game; a district is one of the few purchases you can never refund.
High adjacency pays twice in this ruleset: once as raw science, and again later because the strongest science policy cards scale with adjacency and with how large the Campus city has grown. A low-adjacency Campus in a stunted city and a high-adjacency one in a tall city are not a few points apart in practice — by the Modern era they are different classes of district, and only one of them earns its slot in your policy math.
Let this reorder your settling priorities. A spot with fresh water and a mountain-and-reef pocket beats a slightly richer food tile every time for a designated science city, because the district you place there compounds for two hundred turns. Scouts are not just looking for land; on a science plan they are shopping for Campus real estate, and a settle delayed for a better site is usually worth it.
Choosing and Growing the Spaceport City Early
The Spaceport arrives with Rocketry in the Atomic era, but the city that hosts it should be chosen far earlier. Look for high base production — hills to mine, a strong Industrial Zone position near an Aqueduct or Dam — plus room to grow, distance from hostile borders, and a spare flat tile, since the district demands flat ground. It does not count against the population-based district limit, so the slot is never the problem.
Then develop it like the plan depends on it, because it does. Industrial Zone first among its districts, Factory and later a Power Plant inside, domestic trade routes pointed at it while young, and Magnus settled early so his chop bonus is ready. Crucially, do not chop everything for tempo: woods and rainforest near this city are a savings account, because chop yields scale as the game advances and cash out biggest during the projects themselves.
Consider a second spaceport city if your production allows. The launch projects run in sequence, but the finish line in Gathering Storm is a journey you accelerate with repeatable laser-station projects, and each needs a Spaceport to run in. Two developed spaceport cities let one handle the next launch while the other stacks laser stations — often several turns saved at the exact moment rivals are closing.
Trade Routes and Governors, Assigned by Role
Trade routes are role support, not free money. While your science cities and the production core are young, run domestic routes into them: food and production compress the awkward decades in which a planned district sits in a city too small to place it. Once the core matures, swap toward international routes for gold — by the late game the bottleneck is usually cash for buildings, tiles, and upgrades, since your hammers are reserved for projects.
Governors follow the same logic. Pingala is the science governor and his final promotion accelerates space projects, so the cleanest setup is growing him in a city that is both your best Campus and your spaceport candidate; if they differ, plan his late-game move in advance. Magnus lives in the production core for the chop harvest. The remaining governors patch whatever the plan lacks — loyalty on the frontier, gold in the trade hub — rather than chasing per-city optimization.
Era Checkpoints: What Should Be True, and When
Checkpoints turn a vague sense of "doing fine" into something you can audit. The list assumes standard speed and the difficulty band above; treat each item as a question asked at the era transition, and two failed answers in one era as an alarm, not a shrug. The tech tree checklist on this site helps here — ticking off eurekas along the Education, Industrialization, and Rocketry lines shows which boosts you can still realistically trigger before researching past them.
When a checkpoint fails, fix that axis specifically. Behind on Campus count, settle or buy your way to the missing site instead of vaguely "building more science." Behind on core production, that is the signal to choose the Industrial Zone over yet another Campus — the comparison the next section walks through in detail. Checkpoints exist to change decisions, not to grade them after the fact.
- End of Medieval: three to four Campuses placed in the science cities, Libraries done and the first Universities underway off Education; the spaceport city chosen, its Industrial Zone placed or reserved via Apprenticeship; every trade route slot in use.
- End of Industrial: Universities finished across the science cities, the core's Factory built and its regional reach covering neighbors, expansion essentially done, and your tech count at least even with the strongest AI you have met.
- End of Modern: Research Labs coming online with power to feed them, Rocketry done or a few turns out, the Spaceport queued or standing, and a gold reserve ready to buy the last science buildings.
Worked Comparison: Fourth Campus or Industrial Zone in the Core?
Set the scene precisely: standard speed, early Renaissance, and the city in question is your designated production core. It could place a Campus at modest adjacency, or an Industrial Zone next to its Aqueduct and a cluster of mines. You already own three good Campuses elsewhere. Which district earns the slot? Both are affordable; the question is which one this city, in this role, should want.
The Campus route buys a trickle: the district, then a Library, then a University, each competing for the same production queue, delivering science your three better Campuses already produce more cheaply. The Industrial Zone route compounds instead. Its Workshop and Factory speed up everything this city builds afterward — including, eventually, every space project — and the Factory's regional reach shares part of that gain with neighbors. In the core, under these conditions, the Industrial Zone wins and it is not close.
The verdict flips only in two cases: the Campus site is exceptional — mountains and a reef stacking it well above your existing three — or this city was never really your core and you are fixing the role sheet, not the build order. The transferable principle: in a role-assigned empire, a district is judged by what its city is for, not by what it yields on paper.
Diagnosing a Stalled Science Game — and the Realistic Exits
Science plans rarely die loudly; they drift. The honest tell-tales, checked around the Renaissance-to-Industrial transition: the tech gap to the leading AI is widening rather than stable; your best city would need thirty-plus turns for a space-scale project; every spare hammer goes to units because a neighbor keeps testing you; or your culture is so thin that key economic policies and governments arrive an era late. Any one is survivable. Two or more, worsening, means the original timeline is gone.
The exits depend on which half of the game you still own. Strong production but weak science points at a targeted war: capturing a neighbor's developed Campus cities buys research you failed to build, and spies stealing tech boosts close the gap. Strong gold and alliances but weak hammers point to the diplomatic track instead. A genuine culture base can pivot to tourism. What never works is changing nothing and hoping the AI fumbles; at this band it fumbles the last mile, not the middle game.
Also learn to spot the deficit that merely looks fatal. Trailing in techs while holding a decisive production core is often still a winning position, because AI empires research fast but convert slowly. If your core can finish projects in single-digit turns once they unlock, staying the course beats a panicked pivot — the checkpoint that matters most is the last one.
The Last Fifty Turns: Protect, Convert, and Don't Fumble
Once the first rocket flies, you become the board's problem to solve, and the Spaceport is the obvious target. Station a counterspy on the district permanently — a successful sabotage costs many turns of repairs at the worst moment — and keep walls, a garrison, and a reaction force near any spaceport city a rival can reach. An interior city chosen in the Medieval era pays its final dividend here: geography is the cheapest defense you will ever buy.
Then convert everything into project turns. Cash out the chop bank through Magnus. Finish the Royal Society in the Government Plaza so Builder charges pour directly into projects, and keep buying Builders with gold while that trade compresses the clock. Keep the spaceport city fully powered, hard-buy anything else it wants, and if you built a second spaceport, run laser stations there in parallel with the launches. Every hammer spent on anything else now is a hammer the runner-up thanks you for; the remaining losses come from a short list of unforced errors.
- Building the Spaceport in the highest-science city instead of the highest-production one — everything the district does is a production problem.
- Finishing every science building empire-wide while the first launch project sits idle; past a point, the marginal beaker no longer beats the marginal project turn.
- Ignoring power, so Research Labs and the late core quietly run below capacity for the entire endgame.
- Deploying a counterspy only after the first sabotage — insurance bought after the fire.
- Chopping the core's woods in the mid game for tempo that late-game chops would have repaid several times over during the projects themselves.
- Neglecting culture entirely, then reaching the space race without the government and policy slots the race assumes you have.
Last reviewed: 2026-08-20 · Based on game-data snapshot gs-1.0.12.68-build15296837