The “plant” was cheap. A tent and some leftover assembly line components. Probably their production costs won’t go down as much as they’d like, due to an increase in manual labor–but that’s not the end of the world.
Really, though, we probably won’t know much until Q3. Unless Musk spills some beans at the Q2 call in a couple of weeks. Q2 won’t be profitable but it’s possible they’ll have some extrapolation from the end of the quarter.
Because he’s genuinely smart. And his business accomplishments are from leading a team who does real useful innovation. Trump appears not to even know what he doesn’t know and his business accomplishments seem to be based around squeezing his creditors and money laundering.
The EPA ratings for the Model 3 AWD and Performance version are out. It looks like the AWD version does have a shorter range even if the official range of all versions is 310 miles. The Long Range rear-wheel driver version is probably underrated at 310 miles.
The interesting news to me is the output specifications on the motors (the Specs tab in the link above).
[ul]
[li]LR rear-wheel drive 26kWh/100 miles, 211kW ~ 283 horsepower[/li][li]AWD dual motor 29kWh/100 miles, 144kW + 188kW ~ 445 horsepower[/li][li]Performance 29kWh/100 miles, 144kW + 211kW ~ 476 horsepower[/li][/ul]
I’ve got to say, I’m feeling pretty good about that extra $4000 for dual motor, even if I am getting a rear motor from the reject pile.
That makes me wonder–do I have a binned motor with my LR? The rears on the LR and Performance are both 211 kW; it’s just the AWD that’s lower at 188. Maybe they’ve been stockpiling the lower-perf motors for use with AWD. Our company sometimes uses a similar strategy with computer chips…
I definitely get a healthy 300+ mi even at 75 mph. I’ll bet I’d get well over 310 if I stuck to 65 mph, not to mention driving during the day.
That’s not leftover tooling; that’s all new. I see their tools on shipping floors of our mutual suppliers all the time. The tent is ingenious, though.
I am almost certain you can’t bin electric motors. The capacity depends simulataneously on the amount of wire and magnet sizes. Most likely the smaller one is using the same housing but has less windings in it or some other simple mechanical difference.
It is quite possible the hardware is identical and the only difference is software.
Unlike chips, there are not nanoscale defects - any motor that passes qa should be within a percent or so of any other.
For the pendants : I am saying for a given electric motor topology and voltage, sustained power levels more or less scale up and down with more permanent magnet or induction coil volume and basically just more copper in the windings.
Sorry, I should have said “drive units”–that includes the motor, but more importantly the power electronics as well.
That said, there are a few things in the motor that could cause small variations. Neodymium magnets in particular come in different grades. But these are probably minor effects.
Also, “binning” sometimes just means that it’s what you qualified a part to run at. It may be that every part could run at the higher levels, but if you only test some parts at that higher performance level, and the remainder at some lower level, then you’ve effectively binned them. Granted, this works against any idea of stockpiling lower-rated motors.
Musk wasn’t quite clear on what he meant (and probably exaggerated a bit), but the impression I got is that the tooling was new, but otherwise unused. It’s not exactly a good thing that they had so much tooling sitting around, though maybe it was just stuff that they’d expected to have used already, but hadn’t due to the slowish ramp.
Any thoughts on the reliability of teardown firms? We have two different reports on Model 3 costs that came to basically the same conclusion. If they’re right, it’s a good sign for Tesla. But the skeptic in me wonders if these types of companies are considered reliable. Hard to tell from the outside.
Regardless of the teardown valuation I think the 2170 battery is the item that could make (or break) the uniqueness of the car. It gives them an edge that’s going to fade fast when other manufacturers decide to ramp up.
I considered the combined computer core a deficit and not a bonus. If ANYTHING on that board fails you’re in for an expensive repair.
It’s not just the 2170. It’s the full pack integration where Tesla is ahead of the curve. Obviously, Tesla has to keep moving and they won’t have a lead forever. But I suspect they’ll have a 2-year lead for the next 5 years or so.
Modern electronics are incredibly reliable despite the high parts count. Despite having thousands or even tens of thousands of individual components, a circuit board the size of a sheet of paper can be 100% reliable for decades. And this is an area where the state of the art in manufacturing is already “robots that move faster than you can see”. Every manufacturer has access to this stuff of course, but Tesla is ahead of the game in design.
This is one place where vertical integration is a huge benefit. It’s also a place where I have direct (though limited) experience. Our cubesat project went through a couple of design phases, with the initial expectation being that we could plug together off-the-shelf modules and get a decent end product. We couldn’t. The modules never quite did what we wanted, so we had all kinds of little extra pieces to cover the deficiencies, and the cabling was a mess, and overall it sucked. Eventually we decided to just make our own single board that did everything. It worked fantastically, because it did everything we wanted and nothing we didn’t, and there was no rat’s nest of wiring or anything like that.
We pulled it off because we had the design chops for a custom PCB. Other cubesat teams frequently didn’t have that expertise.
Now, clearly other auto manufacturers have some level of electronics expertise, but it seems like they’re unable or unwilling to integrate at the level Tesla is at. If the teardown people are right, this is a major factor why Tesla is the only company with EV designs that are fundamentally profitable.
Something like that you don’t repair. You replace. Whether this is expensive or not actually depends on the reliability. If the reliability is high, then the supply of those boards on the used market will be higher than demand. This is true for the Prius - it’s computers and inverters cost next to nothing on the used market, even for ones out of low mileage vehicles. This is because Prii get totaled in crashes at a much higher rate than these modules actually fail. (most totaled vehicles have at least 80% or more of their parts in perfectly good condition, they are just too expensive to repair compared to their used market value)
If this board is unreliable, yeah, that’ll suck, though Tesla does offer a 100k mile/8 year powertrain warranty, which should include that.
Generally they’re okay, but we do our own teardowns. However, aren’t these two reports from the same company, and the first one was blistering, and the second one was favorable and recanted the first report?
no, the first one looked at design/engineering and guessed at costs, the second one actually looked at costs. As usual, the automotive media goes with the clickbaity headline and doesn’t understand what’s actually being said.
the second report didn’t change their assessment that the Model 3 is built like crap, but did conclude that on a strictly operational basis the car costs less to build than it does to sell (at their current pricing.)
There are two different companies here: Munro and Associates (that did the recent update as well as an older one), and some German firm. The German one came up with costs of $28k for the LR model based on a 10k/wk production rate. M&A had a figure of 36% gross margins, which translates to $31k. Close enough for government work.
jz78817 is right that their second report (really, a progress update on an ongoing report) didn’t exactly contradict the first. The first one seemed to use a few superficial aspects (adhesive use, sheet metal stampings) as a proxy for costs and Munro wasn’t that positive at the time. The update came after a full part-by-part cost analysis. Tesla’s focus on integration here paid off but it wasn’t apparent except by close inspection.
The report didn’t look into improvement over time–it was based only on a fairly early production model. It’s fairly obvious that Tesla has massively improved their build quality over time. No one is complaining about panel gaps these days.
It’s not really the media at fault here. Munro himself had a very distinct attitude shift and said things like “we’re eating a lot of crow around here.” It’s true that M&A’s conclusion on build quality didn’t change, but the “shocking” development wasn’t the media’s invention; it was how M&A presented their findings.
Sure, the media has some responsibility for putting things in their proper context, but given that they’ve never done so previously, why start now…?
The German report (from a few months ago). One report may or not be believable, but two reports from different firms and fairly consistent conclusions seems hard to contradict.
Wow, the Tesla marketing blitz has begun. I’ve gotten two emails from Tesla showroom reps in the last two weeks, and a voicemail from another asking if I want to come in and test drive a Model S and see if I want to upgrade my reservation.
This is the sort of sales techniques I’d chalk up to a traditional auto dealership. And still no $35k Model 3 that hundreds of thousands of people are waiting for.
I’ve been following this thread mostly as a lurker as someone that’s an auto buff/salesperson (though not for Tesla, obviously) and an on again, off again admirer of Musk, especially as it pertains to SpaceX.
So I just saw this article on CNN.com whereby an analyst from Needham & Co. named Rajvindra Gill has stated that as of right now refunds are outpacing deposits for the Model 3, citing the expiration of the $7500 tax credit and long wait times, along with continued unavailability of the base car (which this person also claims likely won’t see production until 2020).
Tesla of course disputes all these findings, which I guess they always do.
Article here: 24% of Tesla Model 3 orders have been canceled, analyst says
Gill kinda seems like a moron. In particular, his prediction that Tesla won’t manage to build 100k Model 3’s this year is nonsense.
According to third-party trackers, Tesla has already built 52k Model 3s, most of them this year, and is sustaining 4300/wk. There are, what, around 24 weeks left this year? Even if Tesla stops improving, they will build an additional 100k in the second half of this year. There would have to be some great disaster for them to only produce 100k total 3s this year.
The deposit refunds don’t seem like a huge concern. The right number for deposit count is zero. Tesla is leaving their “anti-selling” phase (hinted at by Ravenman’s experience), and may at some point have to work hard at keeping sales up (who knows–they may even need advertisements at some point), but for now they have it easy.