Wouldn't the C rate of the cells matter? If the 2170 can discharge at a higher rate (without damage) than the 18650's then it should be able to generate a similar amount of power even though the pack is smaller. AFAIK we haven't seen any data on the C rate for the gigafactory 2170's but maybe I missed it.
Yes the P100DL has higher performance than the 85 and 90 because they all use the same cells so they're up against the same C rate limit. Once there you have to add cells to get more power.
I'm not being obtuse, I'm actually trying to learn something at this point
Gah, you all are harshing my mellow and now I'm wavering on my Model 3 reservation - may keep driving my Roadster v2.5 until Roadster v4 comes along....
Even if the C rate is higher, it has to make up for the smaller pack size. To get the same C rate out of an 80 KWh pack as a 100 KWh pack, the C rate needs to increase 25% (1/0.80 *100 = 125%). There are some Li-ion chemistries that allow very high discharge rates, but they have other draw backs. Tesla uses the NCA chemistry, which has the highest energy density of any chemistry. They are probably up around 300 Wh/Kg now (the theoretical limit). NCA cells are limited to 1C discharge.
NMC is also used for EVs and Hybrids and can be formulated to do 2C discharge rates, but the best energy density is around 220 Wh/Kg, most cells. The source below doesn't indicate what trade offs are necessary for the higher C, but I wouldn't be surprised if those cells have lower densities (150 Wh/Kg). With battery tech, there are always heavy trade offs.
LFP can ger very high C rates, but have only 1/3 the density of NCA. One of the best chemistries for high C is LTO chemistry, but it tops out at 80 Wh/Kg. It's good for stationary storage where you can have large battery arrays, but would be a poor choice for a car.
Ultimately what is going to sell cars is the best possible range. The car enthusiasts pay attention to the top end performance, but the bulk of buyers are going to be perfectly happy with the performance of the standard model. The performance of a Model S original 60 KWh pack is better than 95% of cars on the road. When comparing the acceleration of the old 60 to other sedans, it's much better than most. The 0-60 of the old S was around 5.5s. A Toyota Avalon is around 6.5s, most Camrys are around 8s, most Ford Fusions are in the 6.5s range, many of the BMW 3 Series are over 6s, etc.
The biggest limitations EVs have is range and price. The price issue is narrowing faster than the range issue. Most buyers of EVs are going to be concerned about the range rather than performance of the most expensive version because performance of whatever they buy (if the car was designed right) is going to blow away anything in the same market niche for performance.
Tesla is going to optimize the 2170s to get the best possible range rather than a higher C rate. If someone stumbles upon a new chemistry which meets all needs (safety, life time, etc.) and happens to have a higher C rate, Tesla won't hesitate to standardize on it, but right now improving range and cost are mission #1 and #2 for battery chemistry. Tesla's engineers will work magic to get the highest performance version to be as much of a super car as possible, but they aren't going to get a special battery for that use if it means compromising in any other area.
A link describing the main li-ion chemistries in use today:
Types of Lithium-ion Batteries – Battery University