Listen, if charging speeds for S3X are to be tripled (per Elon), the cells as used in Model 3 are either going to be done away with, or are just not being utilized to full potential yet (I doubt the latter).
So a 100kWh Roadster pack with 3x as high charging rate, likely will also offer faster discharge rate without overheating issues. The spoils of lower internal resistance chmistries. Rimac gets 1.4MW+ from 120kWh, that's outcome of their mass vs power vs range dilemma, and they are not really bound to a specific cell tech.
Roadster is more expensive than most other cars, so could have externally purchased cells or even pouches. In a $200K+ car, 200kWh will be $20K at the pack level by then again, per Elon). Add a bit for more carbon in stead of aluminum, or higher grade titanium. The battery, if 2170's from Model 3, will NOT be the most expensive part of the car. Or, the car will simply be built for $60K. A more bare Model 3 with more batteries. Fewer seat, windows, window motors, etc, etc.
Musk was talking long term not 2020 and it's unclear if he was including pack size increases or not
Then, Musk never said that any existing vehicles will be able to take advantage of Supercharger V3. It's safe to assume that at least the new cells can take advantage to some degree but don't assume 250kW at peak.
250kW because that's what Musk mentioned as making sense and since V3 arrives late this year at the soonest and they accelerate expansion afterwards, it's safe to assume that the Roadster with a 200kWh pack, won't be able to do better.
That doesn't suggest major chemistry changes enabling much faster charging and we know Tesla is more focused on cost than charging rates. Another way to look at it, much faster charging would come at a higher cost and lower energy density so they physically couldn't fit a 200kWh pack in the Roadster.
Manufacturing costs have little to do with retail price, even more so for low volume products where you need much higher gross margins to recover R&D. Capital intensity is another factor here, low volume can be done without spending much capital and Tesla needs that.
Rimac wants to sell their cooling technology and their cars are showcasing that technology. Spending $, volume, weight, drag on cooling is not always the best solution. They do claim 0 to 80% in 30 mins so not that fast. The C Two can only do 2 laps at Nurburgring.
Model S peak battery power is 451kW with 100kWh of older cells. The Roadster will need 900-1200kW and Model S doesn't sustain 451 for long. Granted there are a few variables here, cooling efficiency, cooling the motors (with S those heat up first), air intake size , speed per cooling needs and ofc they have the new cells. All in all, even with a a 200kWh pack ,it's not gonna be easy.
A 100kWh version is not gonna happen. They kill perf, range , charging miles per min added, would destroy the balance of the car and they would need to redesign suspension, brakes and wheel alignment +software. And then they sell this for 180k? There wouldn't be demand, the investment would not be recovered, makes no sense. Wouldn't even have better handling since you trade weight for higher center of gravity but the track and wheelbase are not changing.
Weight is an issue when you have constraints like in racing where the tire and wheel dimensions are predetermined. In the real world, you deal with it and everything is a trade-off. And BTW, weight being high is just an alarmist assumption.