TSLA Market Action: 2018 Investor Roundtable

[Fact Checking]

EDIT: Nevermind, continued reading and saw it was Vicki. One quick caveat to her numbers: they may apply to the paint shop only. The process could still easily be constrained by cell / pack construction or other points in the assembly process.

I find it hard to believe that this would apply to paint shop only. What's the use of painting thousands of extra bodies? They don't have space at the factory to store a lot of painted but otherwise unfinished bodies.

Correct - once a body is in the paint shop the final assembly clock is ticking and within about ~48 hours a new Model 3, Model X or Model S leaves Fremont.

They have 'overflow lines' of a few dozen cars max to increase buffering and to handle exceptions - but I don't think they can store even 1-2 hours worth of production internally, let alone 1 day worth of cars or more ...

So Vicky's numbers are the real deal, 7.5k/week peak production confirmed.

The only caveat is that these are likely all batches of Medium Range models destined for the west coast, with batched colors and batched options. But that's OK and expected so late in the quarter.

 

[Fact Checking]

The last few quarters they do burst production at the end and then drop off massively at the beginning of the next quarter.

Actually, the end of quarter drop-off of production is regular end of quarter inventory management. BMW does something similar too:

See those big 'waves' in BMW's numbers? Those are the quarterly peaks - they are entirely normal and optimize the capital used at the end of financial quarters. The reason is that they intentionally ramp down the supply chain at the end of the quarter and try to 'drain' existing inventory - this minimizes inventory of parts and finished goods at the end of the quarter.

The consequence of this to Tesla is (and likely to BMW too), that at the beginning of the quarter there's fewer parts in the supply chain pipeline and fewer parts in inventory, and production cannot yet ramp up to full speed. This is done in the first 2-3 weeks of the quarter.

Then slowly ramp back up a little, but not quite to the peak number.

Again, as I said that's false - and you cite no numbers.

I believe if you cited numbers they wouldn't support your argument - in 2017/2018 as the Model 3 was ramping up Tesla has consistently sustained 'peak' in the next 1-2 quarters:

      Quarter   Production         Peak rate           Full quarter rate
      2017/Q4        2,425             1,000                    185/week
      2018/Q1        9,766             2,000                    760/week
      2018/Q2       28,578             5,000                  2,198/week
      2018/Q3       53,239             6,000                  4,050/week
      2018/Q4       65,000 (est.)      7,500 (est.)           5,000/week (est.)
      2019/Q1       78,000 (est.)                             6,000/week (est.)

As you can see, ever since 2017/Q4 Tesla was reliably able to hit their previous 'peak' manufacturing rates of the Model 3 and sustain them within 1-2 quarters, and it's no surprise why that is so: the main limit to sustaining a peak rate is the logistics of the supply chain - with more than 1+ month shipping time of some of the key parts Tesla first has to stress-test their production to make sure they can then consume the parts that get sent.

I took the liberty to estimate 2019/Q1 Model 3 production numbers based on the peak numbers we have from Q3 and Q4, which I'm sure many are curious about.

Q2/2019 should be one step closer to 10k/week with the introduction of the Standard Range model - and Q3 and Q4 of 2019 could show further growth, as the Shanghai Gigafactory comes online and starts ramping up to their planned 3,000/week rate - which will be added to the Fremont numbers.

 

[Fact Checking]

Vincent on Twitter​

"Chinese Media: After Tesla China announced the new pricing structure of Model S and X, many potential customers rush to their nearest Tesla stores to check out the vehicles, many of them were ordered right at the stores, demand largely increased."​

Vincent's reporting about events in China has been pretty reliable in 2018.

 

[Zaxxon]

Except NASA uses the metric system

I was unaware that Liberia and Myanmar had landed on the moon...

 

[tsunamiofhurt]

Vicki is a reliable source and I assume her comment has been authorized. Elon is following her on Twitter.

She is an intentional leak I believe. Nothing spectacular just information Elon wants to get in the market...

Ed McCabe‏ @eddiemac3356
Replying to @VickiSalvador
How many cars this week Vicki?


1h1 hour ago
Replying to @eddiemac3356
A lot ! We’re consistently hitting 1300+ a day. Getting close to that 1400 mark.

Ed McCabe on Twitter

Elon is not following her. IMO there is no way Tesla would randomly authorize certain factory workers to leak production numbers on Twitter. As much as I like to see these data points, she is probably endangering her job.

Looked through Vickies tweets and also found this:



This sounds like a competition between the different stations of production, as she described before. My speculation - find out who can do the equivalent of 10,000/week?!

VICKI SALVADOR on Twitter

In the past she has talked about competitions between different shifts on paint.

 

[Cherry Wine]

I find it hard to believe that this would apply to paint shop only. What's the use of painting thousands of extra bodies? They don't have space at the factory to store a lot of painted but otherwise unfinished bodies.

They could be using backstock of packs already produced for final assembly, could they not? Don’t get me wrong, would love for them to be producing 7500/week, just throwing a dash of caution into the mix.

 

[Artful Dodger]

Can you be more specific on the sourcing of this?

EDIT: Nevermind, continued reading and saw it was Vicki. One quick caveat to her numbers: they may apply to the paint shop only. The process could still easily be constrained by cell / pack construction or other points in the assembly process.

Not EASILY. Paint is in the main production sequence. Everything done to the body before that step funnels through Paint. Everything done afterward waits for Paint.

If you get the sequencing wrong, there is no extra room to stockpile painted bodies. The line slows down instead. Bty packs and drive components for the Model 3 are the exception. They all come from GF1/Sparks and have pre-planned logistics.

So if Paint is nearing 1,400 per day (inc'd 300 S/X) that will be 1,100 Model 3s per day. 7,700 week, and an instantaneous annual rate of ~370K Model 3s and 100K S/X per yr.

That's getting near full production speed for Fremont. Then they'll use the rest of the year to improve quality, decrease rework and overtime, thus increasing margins.

GF3/Shanghai adds another 5K Model 3s per week beginning in Jan 2020. That's approx. 13K/wk a year from now, which is an 85% increase YoY.

So, not easy, but I like it very much!

CHE3RS!

 

[anthonyj]

MACD on the 1 month chart just golden crossed. Hasn’t happened in two years. This week will be huge

 

[UnknownSoldier]

Any news on Europe GF recently?

 

[Esme Es Mejor]

Weekend OT:

How does Hyundai sell this for $37k? Selling at a loss or break even/?

I’m about 10 pages behind, so forgive me if this is all redundant.

I don’t see any reason Hyundai can’t sell the Kona profitably at $37k. The base model gas version starts at $19.9k. That leaves over $17k for the battery without factoring any savings from not needing an engine, transmission, emission controls, etc.

I suspect what limits their volume isn’t profitability, but lack of battery supply. Sounds like they’re limited to less than 40k batteries per year between Hyundai & Kia. Until LG or SK Innovation can be convinced to build a Gigafactory of their own, none of their automakers are going to have mass-market volumes.

The Kona EV looks like a very good car at a nice price, and I hope it does well, but it looks like 4-5 years away from mass volumes at minimum.

 

[Sean Wagner]

Jay Leno gets Tesla.
Excellent article on Clean Technica here:
Jay Leno Wonders Why More People Don't Celebrate Tesla | CleanTechnica

Cut to Jay Leno's video here:

Flashback to 2012 with Jay Leno interviewing Franz with the MS. Shows just how perceptive Jay Leno has been.

...

Jay Leno in 2012 [video above @ 11:51]:
"Range is not really an issue in this car. [Pause] The problem of the electric car has been solved!"

 

[tsunamiofhurt]

Oh my god. You... you mean self-described TSLA bears have a lower estimate of Tesla sales than a pro-EV blog?!?!?!? How can this be?!?!?

Edit: Ah, I see, they’re trying to claim there’s an abnormal lag between delivery and their own estimate of VINs registered there. A database that’s known to lag behind actual changes. Oh noes!

FWIW their estimate of VINs in NMVTIS is close.

 

[sparcs]

I believe if you cited numbers they wouldn't support your argument - in 2017/2018 as the Model 3 was ramping up Tesla has consistently sustained 'peak' in the next 1-2 quarters:

If you look at your numbers though, they are all averaged out. There's no granularity to figure out what they are producing each of those weeks in a quarter.

There have been articles for this quarter showing periods or production between 4 and 5k. I just think it will average out well with the burst production.

like this: Tesla is now able to maintain production of close to 1,000 vehicles per day
and
Tesla Model 3 weekly production slightly down, overall production at close to 25,000 cars this quarter

I have a different perspective that the lows and highs will average out for another good profitable quarter, but they're still having a lot of bottlenecks to work through that aren't making for easy production.

 

[Sean Wagner]

And my point is the Germans don't have to be perfect.

They don't even have to be as good as Tesla.

They just have to beat most of the legacy incumbents and the EV startups.

I don't see that as being incredibly difficult. Tesla being first, having the backing of all the hardcore EV enthusiast, having Supercharger network et al has not had an easy time. New startups have to convince customers to pick their car over a proven Tesla vehicle. Something startup Tesla did not have to do.

IF VW Group is better than GM and Toyota they are in a decent spot.

Precisely. Well, a relatively decent spot, as they survive and get to use their newly honed skills another day.

Ferdinand Dudenhöffer, one of Germany's foremost independent auto experts who famously predicted the rise of Tesla some years back, thinks VW could surpass Tesla in Europe sometime in 2021. From memory - currently missing the link.

Not to be omitted as an important puzzle piece is how many of the world's suppliers big and medium-sized are gearing up for the EV future. They plan on capturing a lot of the value currently accruing to fully-fledged car companies. This also will and does lower the barriers to entry.

Sooner or later, cheap, good-enough Chinese-made EVs will start washing up on foreign shores. The alien dreadnought with some helpful bioware input had better be running by then.

 

[KarenRei]

That's getting near full production speed for Fremont. Then they'll use the rest of the year to improve quality, decrease rework and overtime, thus increasing margins.

Nobody seems to be discussing the fact that Vicki mentioned that the reason they've been getting such a high rate through the paint shop is that they've apparently gotten it so that the paint comes through right the first time and now it's rare for vehicles to need paint rework.

Which means this is a direct reduction in depreciation and labour per vehicle. Excellent news for margins Of course, as for Q4, it'll only have an effect on the margin in the last ~4 weeks, minus any holiday shutdowns.

 

[AlexS]

BUT BUT its so much better for students to practice their math using the imperial system.
Start out with acceleration given as 2 inches per second and an initial velocity of 0.25 yards per minute. Find velocity in MPH at 30 minutes.

That's why my 8th grade math teacher didn't like giving problems with metric units.
PS no calculators then and slide rulers were not allowed on tests.

wait wait, there are no time units in imperial system only metric system has time unit.

So there is no time in United States. Please inform TSLAQ and FlatEarthers(maybe they are the same) about that.

EDIT: my bad, I was wrong, there are time units in imperial system

 

[mongo]

wait wait, there is no time units in imperial system only metric system has time unit.

So there is no time in United States. Please inform TSLAQ and FlatEarthers(maybe they are the same) about that.

Um, we also have the second, along with minutes, hours, days, years (no 10 based values here)
Metric has subdivisions of a second, killoseconds, mega seconds (which is fine when doing programming)

Then there is the degrees vs radians thing (lat and long, elevation and azimuth (no not Isaac) )

 

[BioSehnsucht]

She also said they are getting close to 1,400 in a day (S/X/3). I'm not sure why Tesla would still be pushing towards 8k per week model 3 production if they have exhausted the $45k+ Model 3 order book earlier than planned (as many seem to be speculating). I think the extension of Tesla's order deadline for 2018 delivery could just as well be explained by higher than expected production achieved in the past 3 weeks.
I still think it likely Tesla will shut down Fremont for the last week of the year, potentially preparing the lines for non premium interior and China/Europe production.

I doubt they prepare for non-PUP before Q2. EU/China will happily consume any "excess" for months to come. Possibly they will start building PUP 35K vehicles but that should be not much more of a change in Fremont than adding MR was to the lineup, just a different battery part variation, and maybe a different motor if they determine that adding another motor to the mix will overall reduce costs by building one that is only as capable as needed. The SR complexity will be mostly in GF1 where they will bring up the new SR module / pack assembly line.

But non-PUP will surely wait until after SR is already shipping for some time.

 

[shootformoon]

She also said they are getting close to 1,400 in a day (S/X/3). I'm not sure why Tesla would still be pushing towards 8k per week model 3 production if they have exhausted the $45k+ Model 3 order book earlier than planned (as many seem to be speculating). I think the extension of Tesla's order deadline for 2018 delivery could just as well be explained by higher than expected production achieved in the past 3 weeks.
I still think it likely Tesla will shut down Fremont for the last week of the year, potentially preparing the lines for non premium interior and China/Europe production.

how much modification does the lines need for the China/Euro models? Or only software tweak is needed?

 

[KarenRei]

how much modification does the lines need for the China/Euro models? Or only software tweak is needed?

Different charge port, at least. Probably a list other small differences as well.