Another Look At Tesla Model 3 Mid Range Powertrain And Gross Margin


We dive deeper into the potential powertrain and profit analysis surrounding the all-new mid-range Tesla Model 3.

We’ve had two very good articles on the Tesla Model 3 powertrain and implications on gross margin. Both are based on an analysis by Troy at Teslike (articles here and here).

In those article Troy predicts the size of the battery pack in Model 3 mid range and calculates its effect on gross margin. He predicts a slight reduction in gross margin based on his estimate of the mid range’s battery pack size.

Our results are slightly different. We think Tesla has reduced the pack size more than Troy’s prediction. We also think the mid range’s rear drive unit is lower power than the long-range drive unit. There are secondary savings as well in wiring sizes.

We are predicting a battery size of around 61 kWh usable and a rear drive unit with a horsepower closer the short range’s hp (220 hp).  Therefore, our cost savings associated with going to the mid-range version are slightly higher than Teslike’s prediction.

Our battery is predicted to be approximately 61 kWh usable with a cell count of 3446 cells (vs 4416 in the LR). Seems low? Remember that there is an extra margin in the LR battery pack. Tesla intentionally lowered the range on the long-range version from 334 miles (ref) to 310 miles. In the mid-range version, we are predicting they will eliminate that margin. There are also some secondary effects that result in a pack being slightly smaller: less weight due to the smaller pack increases range, less margin needed since there is no AWD mid-range version.  The mid-range version pack estimate is 17.3 usable kWh’s less than the long-range version and at 115$/kWh that results in a lower cell cost to Tesla of 1990$.

But, the battery is not the only savings in the mid-range version. We estimate the new mid-range motor to have 220 HP (reduced from 271 HP in the long range version). Why such a big reduction? … Slower acceleration and a gear ratio change to help acceleration times since the top speed is lower in the mid-range version.

In addition to the savings in the battery and the motor/inverter, there will be savings in wiring as well due to lower currents.

So, Tesla has dropped the price of the mid-range version $4,000, but they have reduced the cell cost by around $2,000 and moved to a smaller motor, inverter, and wiring (based on our estimate). This should push the cost savings close to $4,000.

*This article was a team effort between the author and Keith Ritter.

Categories: Tesla

Tags: ,

Leave a Reply

36 Comments on "Another Look At Tesla Model 3 Mid Range Powertrain And Gross Margin"

newest oldest most voted

Wait, so part of your hypothesis is no AWD, so no margin needed? Did I miss something? When was it announced that there will never be an AWD mid-range? That seems counter to the direction Tesla has gone with every other vehicle in production…


And this…

@elonmusk For those of us waiting for the standard battery, will we have AWD as an option? #Model3 #Tesla

— Clay McDonald (@Clayticus) August 9, 2018


— Tesla (@Tesla) August 9, 2018


It should be noted though that although SR is supposedly also getting an AWD variant, they did *not* derate the RWD variant in this case… So that doesn’t contradict the assumption that the MR RWD isn’t derated either.


Unless you can cite a source, its extremely unlikely Tesla will build a different motor for the mid-range pack. It is more likely that the reduced voltage caused the lower top speed and current reduction due to a lower C-Rate from less cells causes the slower acceleration. Every extra part they produce will add to cost in organization, production, and inventory. The only change in price is the battery and perhaps modifications to the body/frame that have improved efficiencies in assembly that will effect all models.

philip d

Maybe what they are trying to say is that there will be a total of two motor builds between the SR, MR and LR Model 3s. The SR and MR would likely share the same motor. This motor output would be rated on the top end to fit the maximum output of the MR pack’s size. Then the SR would get the same smaller and lighter motor as the MR but its output would be limited by the smaller C-rate of the smaller SR pack.

Then the LR versions including the Performance Model 3 make use of the second motor build. The AWD and Performance versions for LR then makes use of the front induction motor which is the third motor for the series.


I still don’t think so, but I guess anything is possible. With the limited use of rare earth metals, it wouldn’t be cost effective to stock multiple motors. Plus with how Tesla handles service that would be a lot of extra motors to stock. With other manufactures the dealers own the spares inventory, where Tesla has to carry it on their books. They are supposed to release a AWD version of the standard range, that would further complicate things if they had two low performance motors to save literally a few dollars.

While I doubt it will happen, I’d love to see a Standard Range Performance version. Because with the lighter battery pack, it would make huge improvement to the chassis dynamics.

Reducing the MR drive unit rating to 220 HP reduces cost through more than just the magnets. The other big savings are: – fewer expensive power transistors in the inverter – less copper in the stator windings. – smaller conductors between the pack and drive unit. Tesla has plenty of experience doing this. They designed and built several different drive units for the various Model S and X kWh versions. They would to the same here with the Model 3. Two reasons we believe this is a smaller drive unit: 1. Following on what Phillip D noted above, Tesla likely already has designed, tested, and are tooled up to build the standard-range Model 3 drive unit, which has always been expected to be smaller than the LR 271 HP rear drive unit. Tesla explicitly originally stated the standard range Model 3 RWD was to be 220 HP with 0-60 acceleration at less than 6 seconds. The early versions of the LR RWD owners manual even mistakenly listed the motor as 220 HP. From what we know, the only thing that has held up the SR version release has been that the current LR pack design can not be built in… Read more »
I guess we will have to wait and see, but I really don’t think this is the case. 1. I don’t think Tesla has formally announced that the SR drive unit is different from the MR or LR. Again this would be counter productive to Tesla as they have to carry more load. Plus they have tested their current M3 drive unit to 1 million miles. HP/TQ are a factor of Voltage/Current so any changes to those variables will have a direct correlation, 2. While mathematically the battery might have capacity, that doesn’t relate to the voltage / current rating. I’m sure you know current is directly related to voltage. If the battery pack voltage goes down from fewer serial cells the current draw will increase. I think this is likely due to the decrease in top speed from 145mph to 125mph as most motors have a Volt/Per RPM relationship. In this case the current to deliver the same torque would be increased, leading to increased heat in the wiring, wingdings, IGBTs and batteries. Tesla is known for taking care of their batteries, in order to reduce battery degradation they set a current draw to keep it the same or… Read more »

He already pointed out that according to their calculations, the battery should have spare power, i.e. the only reason for lower power would be a weaker drive unit. (That, or marketing…)

You are probably right that there was not much planning involved — but that doesn’t in any way rule out the possibility that they are using lower-rated drive units that were already in development for the SR variant.

In fact, doesn’t the non-performance AWD LR variant officially have a lower-rated rear drive unit as well? Did anyone ever confirm whether there is an actual difference there?…


Musk said some marketing-y things about P version having “premium bin” components with “double burn-in”. Several people at TMC reported their premium versions showed up to the delivery centers as AWDs, though, and Tesla just loaded new software. So…..


…..So Tesla’s reps took AWD vehicles with drive units that probably did NOT pass the factory double-burn-in tests made software changes that allowed them to operate at the higher amps?

I see “drive unit warranty service” alerts in the future.

Adam – I, too, appreciate intelligent conversation about these topics. I like how you’re working to understand this and that is why I contribute this site – to help others understand the inner workings and to hypothesize on what technical aspects might be working behind the scenes that haven’t yet been revealed. To help you and other readers along with this technical discussion, below is a basic primer on EV battery architectural concepts and specifically, the Model 3 battery packs. Though Tesla made a substantial cell reduction, that reduction did not affect pack voltage at all, as Tesla will be connecting the cells to reduce the number in PARALLEL, not the number in series. That is not conjecture. That is how EV manufacturers do it when they produce packs for the same vehicle with different kWh ratings and cell count. Everything electrical downstream of the pack is designed around the nominal pack voltage. Heaters. AC compressors. onboard chargers. battery BMS. DCDC converters. Etc. That is way I have 100% confidence that the MR model 3 pack will have the exact same voltage as the LR pack. 390 volts DC, using 96 cell groups in-series. That is also how all Model… Read more »

I’ve seen claims that some Model S variants actually use a lower pack voltage? No idea whether there was any truth to that, though.


I dunno….

The LR and LR AWD Performance both use the same high power motor.

The LR AWD uses a lower powered rear motor.

Difficult to imagine a 3rd motor.


Maybe the MR is using the lower-powered rear motor from the LR AWD config?


Exactly my thought.

(Also probably the same one that was intended to be used in the SR variant.)


Given the abrupt way that Tesla introduced the Mid Range Model 3, I think that was rather suddenly decided upon and rushed into production. If so, then there would have been no time to give any new motor type the testing that any such critical part of a mass produced car’s powertrain should get before production.

Using a rushed-into-production new motor type is not something I find believable. What CCIE suggests, that appears reasonable, at least to me.


My guess is they have been working on the motor/inverter/etc for the standard range for some time. I doubt it was rushed. What might have happened is the company decided it had production running smooth enough to sell a lower cost vehicle. I have no idea what this really costs but for sure the margin will be higher than standard pack. It appears end of the year they will change part of the design to be lighter, and between now and febuary its anyone’s guess which motor goes in the rwd standard and mid range.


It was introduced to the public abruptly — but we don’t know how long it has been in development before that…


Yes. See additional comment response below


“Unless you can cite a source, its extremely unlikely Tesla will build a different motor for the mid-range pack.”

I, also, find that rather unlikely. I always appreciate reading the “deep dive” analyses by George and Keith here at InsideEVs, but they seem to have a fondness for jumping from their own speculations to conclusions, with no evidence to support those conclusions.

Bolt driver

I would expect at the volume now, they probably wouldn’t do a separate motor. It seems more of a rush to use less batteries per car. I also wonder on pricing vs cost, the supposed extra 25kwh of battery costs $9k, the guessed size is 12-15kwh less yet the price only went down $4k. Seems like profit might go up unless the $9k had lots of profit and no real cost.

But that is not to say that larger volume might move them to two motors. I know ford did two suspension designs for the ranger trucks, one was torsion bar, and the other coil spring. The coil version was $6 less per truck. At the volume of Rangers and the expected life of the design it was worth doing two designs. Tooling probably ran into the $3-4million range, but at $6 per truck X 500k per year, it paid for itself.


I seriously doubt using a smaller motor or wiring will result in substantial saving, maybe 50 bucks in the long term, definitely not anything close to $2000.


Agree with minimal cost savings. Different equipment sounds more like something Gm would do. They would build a radiator without a freeze plug if it would save then 25 cents.


I don’t know that I’ve ever noticed a freeze plug in a radiator. Blocks, yes. Heads, yes. But never in a radiator.


I think he means coolant drain plug. I’ve run into this on more than one car, where you have to pull the lower radiator hose in order to drain the radiator because there is no drain plug for the radiator. Usually on cars that have supposed “lifetime” coolant.

(Sidenote — when you hear of a car having a “lifetime” fluid, keep in mind that Ford took a case all the way to the US Supreme Court where they claimed that the “lifetime” of a Taurus was 80K miles, so owners out of warranty suing them over various problems had no case because they got the expected lifetime use out of their vehicles. Ford won that case….)


How about they would put a cheaper detent spring in an ignition key cylinder if it could save them 5 cents.


$50 seems a bit on the low side — but I totally agree that the difference is surely nowhere near $2000. Maybe $200 or $300.


3446 cells? That makes no sense. If they reduced the number of cells per group, the closest would be 96 x 36 = 3456. If they reduced the number of cell groups, the closest would be 75 x 46 = 3450. (Though an odd number of groups seems extremely unlikely with the symmetric battery layout they use…)

I think the former is way more likely, since the SR pack was supposed to be 96 x 31 — so they should already have everything in place for shrinking the groups; while reducing number of groups would be a way more substantial change.

(I think though that your estimation for battery size reduction is a bit on the high side… 96 x 37 would actually be my best guess.)


3446 is indeed nonsense. Not divisible any further than 2×1723. I presume it’s a typo and they meant 3456.

3456 is a bit too small for 260 EPA, IMHO. I’d also expect Tesla to leave a little range headroom in case they decide to do a MR-AWD at some point. Pretty sure they won’t do a MR-P, so they don’t need 20+ miles of headroom like the LR.

The weight delta also indicates more than 61 kWh. Especially if they save a few pounds with a smaller motor.

I originally guessed 96 * 38 = 3648 cells for just less than 65 kWh. Tesla always uses 96S, and 38P is a bit below the halfway point between the SR’s rumored 31P and the LR’s 46P. It just feels right to have MR near the mid-point.

$50 is a bit low for the motor savings (assuming there are any motor savings). $100-150 is a decent guess. Power electronics could save $100 or so. Thinner wires? LOL.

I could see adding a few hundred bucks to Troy’s $1409 cost savings. But $4000 is a complete goofball number.


96S36P is correct for what we estimated. Yes, we had a typo in the exact count.

Re: range. George and I have evolved a pretty sophisticated performance dynamics model that pretty much nails every EV we’ve tested it with. The model showed 60 kWh usable for 260 miles EPA range. The debate is how much fixed “buffer” to figure. We have each cell rated at 17.7 Wh/cell from various sources.

Thinner wires? One example would be the charger cable. the MR likely uses 2/0 Cu because the pack will be rated at no more than 109 kW. The LR uses 3/0 per Ingineerix. Copper cable and the connectors can be pricey when it is fatter than your finger.

I have no idea where Tesla came up with $4,000 price cut. unless there is even more de-contenting that they haven’t referenced. Or maybe Panasonic isn’t cutting them such a good deal after all. Or there is a lot more labor to the module and pack assembly with more cells. I believe the packs still have a substantial hand-build aspect to them.


I think it should be obvious that the $4,000 selling price reduction doesn’t correspond to a $4,000 production cost reduction. UBS assumes in their calculations that options in the automotive sector generally have ~50% margins, i.e. $4,000 selling price difference would correspond to about $2,000 manufacturing cost difference — which sounds about right IMHO.


The Mid Range has the same pack size as the Long Range (only fewer cells), so the cost reduction is in the cell only. The quoted number $115/kWh seems high since Musk mentioned a few months ago they were close to $100/kWh cell cost (not sure if it was achieved at this time). Using $100/kWh, the cost reduction would be ~$1700, not bad but not $2000 either.


No, Tesla’s cost reduction would be more than just the cost of the removed cells. Every cell in a Tesla pack has a fusible link (wire) welded to it. So reducing the number of cells will also save on the cost of those wires and that welding, and the labor necessary to assemble those cells and wires into the pack.


With the price $4K less, I would expect that Tesla’s costs are portionately less based on margins. Tesla repeatedly has set prices based primarily on margins, with specific margin targets determining their prices they charge. I expect their costs to be aligned with those known, well publicized, long term margin targets.


Doubt it. The only difference is prob the cell count. Everything else is software limited. Just like the AWD And Performance. No difference in HW