Sorry, it took me a bit to finish reading Ultimate Campaign. All right, a comparison of the kingdom building rules in "Of Cities and Kings" from Rivers Run Red, the second issue of the Kingmaker Pathfinder Adventure Path, and the kingdom building rules from the Pathfinder RPG hardback Ultimate Campaign.
They are basically the same rules, slightly expanded. What is 10 pages in Rivers Run Read is 30 pages in Ultimate Campaign, plus some pages of optional rules.
Some of that is more options - like in terms of Leadership Roles, Rivers Run Red has 11 and Ultimate Campaign has 14 - and some of it is padding; for example the Councilor entry has the same mechanics but 6 lines of fluff verbiage rather than 1. So that additional page count is part more stuffs and part blabbering (though if you were confused by the sparsity of explanation in the Kingmaker version, that's not necessarily bad).
The "Improvement" phase gets renamed the "Edict" phase and other such, but in general you're getting the same system with a bit more of each of the lists-of-bits. There are some more buildings; you can do terrain improvements (waterways, bridges)... What it doesn't appear to have a bit more of is playtesting. Same system, most of the same holes. They did do bolt-on fixes to plug some of the most well known holes, like in the Income phase you can no longer sell magic items created by your city (no real in-game reason given, just to stop the "infinite monies exploit" popular among Kingmakers). They also took the Economy +1 and left only the Loyalty +1 benefit on graveyards, to avoid the Necropolis Gambit. So the various costs and DCs are left 90% the same, with some tweaks.
In general you are getting the exact same system in Ultimate Campaign (same grid layout page, same mechanics) with some additions and some cursory changes to prevent the most well known of the exploits in the previous system.
In the Kingmaker Adventure Path, the PCs start with 50 BPs but this reflects a specific amount bestowed by the swordlords of Restov on the PCs to get them started.
In general, when you start you'd estimate how many BPs you want to grant. Ultimate Campaign describes BP:
Build points are an abstraction representing the kingdom’s expendable
assets, not just gold in the treasury. Build points include raw
materials (such as livestock, lumber, land, seed, and ore), tangible
goods (such as wagons, weapons, and candles), and people (artisans,
laborers, and colonists). Together, these assets represent the labor
and productive output of your citizens.
And also that
In general, 1 BP is worth approximately 4,000 gp; use this value to
get a sense of how costly various kingdom expenditures are.
So in your case... Goblins. The approximate worth of most goblins and goblin settlements frankly approaches zero in raw materials, capital, and skilled labor.
Providing a seed amount of BP at the start of kingdom building means
your kingdom isn’t starving for resources in the initial months.
Whether you acquire these funds on your own or with the help of an inf
luential NPC is decided by the GM, and sets the tone for much of the
campaign.
So your campaign, your vision, but if I were starting out a bunch of goblin PCs trying to wreak a kingdom from a goblin tribe the answer would be "0 BP to start with." Unless the PCs have managed to get enough treasure to cash in for some starting BP.
I'm somewhat concerned by your statement that you'd be "giving them DM-fiat cash just to avoid the kingdom falling appart." You shouldn't bother using a ruleset if the option for failure is not there. If you intend for it to succeed no matter what, then just make it succeed. You should use these rules if failure is an entertaining option (and with Pathfinder goblins, it certainly would be IMO!).
If I were taking We Be Goblins or something and extending it into a kingdom-builder I'd be tempted to reskin the rules and use Junk Points and restate all the buildings and whatnot as the degenerate ridiculous crap goblins would come up with. I bet your players would be super entertained in participating in that process.
Though I think an "Unrest death spiral" is the expected outcome for a goblin city, you could bootstrap this by using the Downtime rules, also from Ultimate Campaign, to do some building and organization development at higher resolution and then once there's some starting points that would be equivalent to BP, move into the larger kingdom building rules. (In fact, here's a sidebar where they discuss the interaction between those two rulesets.)
Best Answer
Note that I have only just glanced over the Kingdom Builder rules, but I see two aspects to this question.
On the one hand, as was stated by SevenSidedDie earlier, which I will just full quote here for simplicity and completeness:
Farms aren't individual farms in that passage, they're a type of change to the hex. You can only add farms to a hex once—after that, the hex has farms.
One the other hand, there's this bit of text in the Watchtower entry:
The fact that it is explicitly forbidden to "stack" watchtowers on a single hex, implies that it is not forbidden to have multiple farm improvements (or roads, highways, fisheries, etc.) on the same Hex.
So as far as I am concerned:
By RAW, you can. The Rules don't mention a limit, even though they explicitly do in a (mostly*) equivalent case.
By RAI, you cannot. The rules not mentioning any kind of limit leaves us with exactly two non-arbitrary** choices: "One", or "None"
However, common sense dictates that there is limit on the number of farms on a single hex, because there is a natural limit to the area, and a less natural limit to the efficiency of your farms (due to technology, magic, amount of slaves you can efficiently throw at it, etc.). Similar logic can be applied to the other starred improvements as well, e.g. once you have turned the whole hex into a parking lot, you can't place any more highways.
*Unlike Farms, Watchtowers grant a one-time bonus as well.
**For any other limit N, you could equally justify N+1.