It requires 10 Uranium Ore instead of 4 Uranium-238, which would potentially be much more efficient since normally you get around 1 u-238 per 10 ore, but in this mod fusion gives unlimited amounts of u-238. The KI2 core consumes uranium constantly regardless of whether you're producing anything, and the ore is a strictly limited resource! Limited ore means you want finite consumption, and aside from infinite research, Factorio has finite consumption.

All said, I'm probably not going to use the effectivity 3 recipe, since I'll have fusion before I unlock it. I'd say it should require 1 uranium-238 instead of 10 ore to be in line with the previous ranks' viability to be run infinitely. Requiring 1 battery along with it could be fine too.

On a side note, the u-238 requirement for running the KI2 core is a bit extreme. Despite using 500MW of my 2x2 uranium fuel rod reactor, my rich water uranium recipe's production of u-235 doesn't produce enough u-238 byproduct to run the KI2 core. I had to put in some circuits to make it produce extra u-235 to keep up with u-238 demands, and my uranium mine can't even keep up! That's bonkers. If I had known that it would eat this much u-238 I would've set up my copper production to use the rich water recipe for the extra u-238 byproduct.

I set up my copper processing to generate the u-238 byproduct and unfortunately the hydroflouric acid consumption is too high for my oil wells to support, despite having about 700% yield. It also generates too much rich water for me to use so the majority is being voided. I think I'm going back to dirty water and leaving the KI2 core offline until I have a uranium-producing star engine up and running.

I just ... mined & centrifuged some uranium the boring old fashioned way, two small patches were more than enough to run three KI2 cores (I wanted more cubes). Voiding rich water sounds painful, that stuff is comparatively tricky to get. The copper processing path struck me as an alternative way to get nuclear online without uranium ore, rather than a primary source for KI2.

"Infinite" isn't really the goal resource profile, since you can't Star Engine petrochem stuff. Plus "infinite" is only ever bounded by finding & connecting more patches anyway, and besides infinite tech, simply defending also requires "infinite" resources.

Tricky? You can just make rich water from coal and stone... Did you miss the Rich Powder recipe?
I only have one uranium ore patch. It's rare! It's also a rather small patch. Usually that's not a problem because uranium is almost entirely used for nuclear energy, which is so efficient you barely need any to power a large base. 248k's KI2 core breaks the mold in two ways: by requiring a massive amount of energy to run and requiring a large amount of u-238.

Rich powder is the comparatively tricky process, yes. Solution canisters take two kinds of acid, several involved resources are both consumed and produced, needing to be balanced / prioritised / voided. It's not Seablock, but it is more than vanilla's Kovarex.

Arguably breaking the mould is a goal of overhaul mods. Compared to the base game, 248k already requires a ton more stone than you expect. I also found that my first KI2 core helped me in not having to warehouse as much u-238 as I usually would.

I only have one uranium ore patch

You know what we're gonna say about not having enough mining. Time to engage the violence machine.

Having just mathed it out (in my head, no guarantees) one KI2 core with effectivity 3 needs 80 uranium in 45s. You'll get this with just 8 drills, no mining productivity. Seems viable?

Yeah I guess I could go claim more territory so I can run another uranium mine to fix the issue, and also get more oil wells. However, the KI2 core isn't actually helping me much? It's 4 modules instead of 2, on a beacon that takes up 50% more space even in an optimal placement scenario... and both the core and the beacon cost 10x more energy. It's not worth it for me right now. I've already produced enough red block research to research stuff that isn't KI2 effectivity so there's no point until I need to start running KI3 beacons along with the awful stokes/gauss modules. By then I'll have star engines and black holes though.

The larger effect radius means the same number of beacons can affect more assemblers. Between that and double effect broadcasts, it can be a big deal if the beacon cap is pressuring you. That's likely an artefact of my 10x tech multiplier though.

Interesting prioritisation on modules vs Star Engines & black holes; I've been leaning a lot on the stokes modules (and KI2 beacons) from way earlier than that. But in retrospect, mostly just smelting & petrochem, things you could absolutely just go mine more and do larger builds instead.

I found two things where stokes modules help:
1, oil processing facilities. Sadly, these are mostly useful for creating acidic water, since kerosene is useless late game and you can't produce only vanilla oil products with the facilities, there's always a significant amount of product related to acidic water or kerosene.
2, hydrofluoric acid. This is the biggest impact. With stokes 3 modules, many uses of hydrofluoric acid become extremely efficient because making the acid creates acidic water which can be turned into sulfuric acid easily, which is the main ingredient of hydrofluoric acid (let's be real here, fluorite and lithium are free if you have any demand on your dirty water ore processing at all). If you then put even the weakest gauss modules on the sulfuric acid recipe, the loop becomes positive!

And as far as the larger radius on beacons is concerned... well, it's actually the same radius. KI2 beacons are just larger, which means the radius extends from more squares. KI1 beacons are 2x2 instead of 3x3. If you stick a KI1 directly next to a machine, you can stack a second one next to that one and have them both reach the machine. That makes it actually better than a KI2 beacon in some situations.

KI2 beacons should be able to use 6 modules right away. I really do not understand why you have to wait until KI3. Hell, once you have access to KI3, you get double module effects and what looks like 20 module capacity on the KI3 beacons, so why would you EVER use KI2 beacons?

It's perhaps not quite that straightforward. The KI3 effect doubling tech needs 100x the research of the rocket silo, with space + exotic science as well. An even larger gap from Star Engine through black / white holes. I'm quite skeptical about pushing through Star Engine, getting the holes, and then deciding it's time to upgrade beacons past KI1.

Overall to me the progression of KI1 -> KI2 with 4 effects -> KI2 with 6 effects -> rebuilding for KI3 effect radius -> KI3 with 12 effects, all those stages seem distinct & well enough paced, but a 10x tech multiplier does stretch everything out.

The Gauss modules are also crazy expensive, acidic water probably one of the last processes I'd socket. You might just be richer than my base is though.

Also, you can do the second-beacon-stacking thing with all beacon tiers. Each effect goes 1 tile further than the size of the beacon. Also, also, you know there's a route from kerosene to rocket fuel?

Oh, I did not notice KI2 beacons reached 4 squares out from the edges. That makes them a lot stronger than I thought. Takes up a lot more room though, and you can't sneak them inside belts as often.
Personally I would not want to modify blueprints to fit in larger beacons. Either plan ahead for them or don't rebuild at all.

I didn't actually put gauss modules on the sulfuric acid recipe, because a positive loop with two outputs has the potential to clog. Also the savings are rather low in the end; the stokes modules in the hydrofluoric acid are enough.
I did put gauss modules in some centrifuges with the "exotic kovarex" recipe, to make exotic matter for free. Lots of beacons stacked on them, makes a ton of pollution but it's a better alternative than using deep space ore from rocket launches.
The entry level gauss modules are pretty expensive but if you just let your module production turn excess modules into gauss modules you'll get a slow trickle which was enough for what I wanted. I'm currently trying to get my hands on enough of the second rank of gauss modules to make stacked deep data more efficient.

Kerosene can build rocket fuel technically but it's not as efficient as using light fuel(especially since you can't use productivity modules with the kerosene to rocket fuel recipe, for some reason), so I didn't see much reason to except to dispose of any excess made to keep up with acid water demand. Then later on I eliminated that problem completely by funneling excess kerosene into the coal driven recipe that turns it all into other stuff. Did you figure out some way to make it efficient, without losing out on petroleum gas byproduct?