Near-future quarterly financial projections

[Doggydogworld]

Are those numbers deliveries or production?
I want to be sure I'm using the right source.

CPCA
January: 59,845
February: 56,515
March: 65,814
April: 10,757
May: 32,165
June: 78,906
July: 28,217
August: 76,965
September: 83,135

Giving totals of:
Q1: 182,174
Q2: 121,828
Q3: 188,317

Thank you!

The CPCA numbers you show are wholesale. That's supposed to be sales within China plus exports. I don't know if a car counts as exported when it arrives at the dock and paperwork is approved, or if the ship has to actually leave the dock.

Wholesale loosely matches production, but not exactly because some cars built at EOQ are not delivered or exported by the stroke of midnight. (And some cars left in that state the prior quarter roll into the current quarter).

Production numbers come out later in the month. They used to come from another agency, CAAM I think, but the numbers I've seen lately come from Gasgoo and/or some guy tweeting (not Moneyball).

 

[ammulder]

As we all know there are more variables than there are data points so a bit of judicious guessing is required to fill in the likely Q3 production numbers.

Here is my updated best stab at it. To be truthful I'm not quite sure which is the right China data to use as production versus sales, but I don't think it matters greatly in the overall scheme of things given the size of the other error terms in model/factory production apportionment.

View attachment 864771

The thing to pay attention to is that either Fremont has unlocked additional production, or that Berlin+Austin have steeper ramp rates than Fremont did. For now I am tentatively assuming that the Fremont ramp rate is more typical than the Shanghai ramp rate. I am tentatively assuming this because Shanghai benefitted greatly from being stuffed by Fremont, and quite a lot of new stuff is happening both at Berlin and Austin. That in turn suggests that Fremont is perhaps increasing production now, up perhaps from 116k to 124k, so an extra 7k of 3/Y (plus an extra 2k of S/Z). I am unsure - there simply isn't enough firm data. In the ramp rate graph below I have dialed in the Q4 as well on a similarly matched trajectory and you'll see it gets to 1,346,910 production for the year.

One can pick holes in this. For example is LFP pack supply to Berlin enough to meet the 15k ramp at Berlin ? Might Austin even be ahead of Berlin and if so is that reliant on Sparks cells, or 4680 cells, or what ? Can Fremont really have increased production by a further 7k of 3/Y in the last quarter ? I don't know the answers to these questions and woukd be grateful if anyone with hard data could help fill in the gaps a bit.


View attachment 864774

Didn’t we just get a report of >80k production for Shanghai in September? (I think Rob Maurer reported that) That makes forecasting 225k for all of Q4 seem pretty low — unless you’re trying to accommodate further lockdowns or something.

 

[petit_bateau]

The CPCA numbers you show are wholesale. That's supposed to be sales within China plus exports. I don't know if a car counts as exported when it arrives at the dock and paperwork is approved, or if the ship has to actually leave the dock.

Wholesale loosely matches production, but not exactly because some cars built at EOQ are not delivered or exported by the stroke of midnight. (And some cars left in that state the prior quarter roll into the current quarter).

Production numbers come out later in the month. They used to come from another agency, CAAM I think, but the numbers I've seen lately come from Gasgoo and/or some guy tweeting (not Moneyball).

Thanks. Yes it is the wholesale numbers I was mostly trying to use whenever I could locate them. Like you I don't know exactly what constitutes an export but at least it is on the conservative side as a proxy for production because a car must have been produced to reach the port's loading dock. I try to use the exact number if I can get it (for personal audit & sanity purposes) and be as consistent as possible (hence wholesale as getting the definitive production ones is very hit-and-miss). I'm not too fussed about hyper-accuracy, for me it is the trend and understanding/explanation that matters most.

Didn’t we just get a report of >80k production for Shanghai in September? (I think Rob Maurer reported that) That makes forecasting 225k for all of Q4 seem pretty low — unless you’re trying to accommodate further lockdowns or something.

Thanks. I'm using 75k/month for Shanghai in Q4. I'd love to be pleasantly surprised by the ramp of the new line but I prefer to be cautious. It doesn't directly affect me to be cautious as I don't trade short term at all. Besides which I think there is too much breathless optimism sprayed around and a dose of reality is a good antidote. Hitting peak production in bursts is different than sustaining it. When I update 2023 I may take a more positive view depending on how the next few months turn out and what (if any) new info comes to light.

==

What is interesting is that nobody has made any significant critique of my Q3 apportionment guesstimate between models & factories. So either nobody has any secret info to share (!) or the gesstimate is roughly on the ball (?). Either way that is interesting because there are some pretty significant implications in that as I set out.

 

[Artful Dodger]

Lol Troy. A couple weeks ago Tesla implied they’re going to shoot for the upper 400s.

I’ll bet you Troy’s initial delivery estimate here is a minimum of 50,000 cars too low.

Indeed. Let's get his P&D forecast into TMC's permanent record for future reference:

Troy Teslike on Twitter: "Here is my first estimate for this quarter as of Oct 9. Tesla Delivery Estimates for Q4 2022: • My estimate: 400K • Analyst consensus: 435K I estimate 454K for production. The large gap between prod and deliveries is related to eliminating the delivery wave in Europe & China. https://t.co/prevQjzM2K" / Twitter​

I *ahem* disagree.

Cheers!

 

[Artful Dodger]

Thanks. I'm using 75k/month for Shanghai in Q4. I'd love to be pleasantly surprised by the ramp of the new line but I prefer to be cautious.

Haha, 'cautious' is a emphamism for 'low'. Even notorious Roto-REUTERS says 20.5K/week at 93% utilization rate:

Exclusive: Tesla to keep output at upgraded Shanghai plant below maximum -sources | REUTERS (2022-09-27)

SHANGHAI, Sept 27 (Reuters) - Tesla (TSLA.O) plans to hold production at its Shanghai plant at about 93% of capacity through the end of year, despite a recent upgrade, two people with knowledge of the matter said, in a rare move for the U.S. maker of electric vehicles.

Since the plant opened in its second largest market in late 2019, Tesla has sought to run the facility in China's commercial hub at full capacity, and recently upgraded its weekly output by 30%, to a maximum of 22,000 vehicles.

So, call it 266K total production at Giga Shanghai in 2020Q4. Also, your proportions for 3 vs Y production in Shanghai are way off. Here's Sep 2022 production numbers for reference: (about 1.7x 'Y' vs '3')

Tesla produced a total of 82,088 vehicles in September which was comprised of 30,257 Model 3 and 51,831 Model Y.

Cheers!

 

[petit_bateau]

Haha, 'cautious' is a emphamism for 'low'. Even notorious Roto-REUTERS says 20.5K/week at 93% utilization rate:

Exclusive: Tesla to keep output at upgraded Shanghai plant below maximum -sources | REUTERS (2022-09-27)



So, call it 266K total production at Giga Shanghai in 2020Q4. Also, your proportions for 3 vs Y production in Shanghai are way off. Here's Sep 2022 production numbers for reference: (about 1.7x 'Y' vs '3')

Tesla produced a total of 82,088 vehicles in September which was comprised of 30,257 Model 3 and 51,831 Model Y.

Cheers!

You may have missed my post on this thread on Sunday where I gave the monthly breakdown of the Chinese numbers I have used, as follows

From China we have approximately the following Shanghai Q3 production :
- 3 = 56,494 = 11261+14976+30257
- Y = 141,458 = 27458+62169+51831

So it seems the 3:Y ratio out of Shanghai went awry in August for whatever reason and that is what is affecting the overall quarterly ratio. I have no insight as to why.

Let's see if they can keep the monthly Shanghai production above 80k. It would be very welcome.

 

[mongo]

You may have missed my post on this thread on Sunday where I gave the monthly breakdown of the Chinese numbers I have used, as follows

From China we have approximately the following Shanghai Q3 production :
- 3 = 56,494 = 11261+14976+30257
- Y = 141,458 = 27458+62169+51831

So it seems the 3:Y ratio out of Shanghai went awry in August for whatever reason and that is what is affecting the overall quarterly ratio. I have no insight as to why.

Let's see if they can keep the monthly Shanghai production above 80k. It would be very welcome.

Oh, I know this one.
Shanghai upgraded the Y line first, so it had higher production over the rest of the quarter, including August when the 3 line was down for upgrades.
July: factory powered down, Y line upgraded
August: 3 line upgraded.
Other news reports say line testing will continue through November, so expect further rate increases.

Tesla China delivers 28.2k vehicles in July amid Giga Shanghai upgrades

Giga shanghai finished upgrading the Model Y assembly line in mid-July. Upgrades to the Model 3 assembly line were expected to finish by August 7, with production restarting on August 8.

 

[Artful Dodger]

So it seems the 3:Y ratio out of Shanghai went awry in August for whatever reason and that is what is affecting the overall quarterly ratio. I have no insight as to why.

The Model 3 line was down for upgrades through at least the first week of August. That was discussed in the main thread. Here's a link to a tweet:

James Cat on Twitter: "Post-update, Giga Shanghai is pumping out 2k Model Y/day. Model 3 line still down for upgrades and will resume on Aug 8. $TSLA https://t.co/7zF6SpwVHi" / Twitter​

Note: he used to post at TMC under the handle "mtlhabsfan" or similar such, but I can't resolve a link to that acct at the moment.

 

[bkp_duke]

The Model 3 line was down for upgrades through at least the first week of August. That was discussed in the main thread. Here's a link to a tweet:

James Cat on Twitter: "Post-update, Giga Shanghai is pumping out 2k Model Y/day. Model 3 line still down for upgrades and will resume on Aug 8. $TSLA https://t.co/7zF6SpwVHi" / Twitter​

Note: he used to post at TMC under the handle "mtlhabsfan" or similar such, but I can't resolve a link to that acct at the moment.

Model 3 lines are also much longer and less efficient than the Model Y lines. Throughput on those is much lower because the design doesn't utilize castings.

 

[heltok]

It seems that a lot of 2019 cars are for sale now:

https://twitter.com/x/status/1582171743816323072

$36k incl FSD, which I assume Tesla added without any cost for themselves. $35k was the original price for SR+ in 2019. I assume that when Tesla did the calculations for leases in 2019, they didn't use 100% resale value. More like 50%? So each resold car is another 17k of profit once they can realize all the FSD revenue? Anyone have figures for how many leases and can do the math of how much profit they will make on reselling these?

 

[thx1139]

Tesla Berlin confirmed a 2000 week on October 1st. So if no further ramp occurred Q4 should be 28K. The plan is to be 5K per week by end of quarter. If we start at 2000 per week 1st week of October and add 250 per week we end up with 5000 per week last week of quarter and a total of 45,500 vehicles produced at Tesla Berlin.

 

[Hytra]

TroyTeslike posted his estimates on Twitter but has already deleted them. From memory, his non-GAAP EPS was at 1.08.
He forecasted a lower automotive gross margin than in Q2.

 

[MP3Mike]

TroyTeslike posted his estimates on Twitter but has already deleted them. From memory, his non-GAAP EPS was at 1.08.
He forecasted a lower automotive gross margin than in Q2.

Yep, here is what was shared:

 

[mongo]

Yep, here is what was shared:

View attachment 865104

Worst automotive GM in a year???

 

[mongo]

Worst automotive GM in a year???

Including the quarter of Covid shutdowns???

 

[Gigapress]

Thanks. Yes it is the wholesale numbers I was mostly trying to use whenever I could locate them. Like you I don't know exactly what constitutes an export but at least it is on the conservative side as a proxy for production because a car must have been produced to reach the port's loading dock. I try to use the exact number if I can get it (for personal audit & sanity purposes) and be as consistent as possible (hence wholesale as getting the definitive production ones is very hit-and-miss). I'm not too fussed about hyper-accuracy, for me it is the trend and understanding/explanation that matters most.



Thanks. I'm using 75k/month for Shanghai in Q4. I'd love to be pleasantly surprised by the ramp of the new line but I prefer to be cautious. It doesn't directly affect me to be cautious as I don't trade short term at all. Besides which I think there is too much breathless optimism sprayed around and a dose of reality is a good antidote. Hitting peak production in bursts is different than sustaining it. When I update 2023 I may take a more positive view depending on how the next few months turn out and what (if any) new info comes to light.

I wouldn't call 75k/month for Shanghai "a dose of reality". Shareloft via Rob Maurer provided a credible report of September production of 82k, so you are predicting that the Q4 average is going to be 8.5% less than the rate already achieved a month ago, when the new production line upgrades had only recently been installed in August and there were likely still some kinks being worked out from the changes.

Your estimates imply a significant reversal of the long-term growth trend, which thus far has been interrupted only by government fiat and by planned production pauses for line upgrades. This would be the first time in the entire history of Shanghai in which a quarter without line stoppages had less output than the 3rd month of the prior quarter.

That prediction could be correct, but please present a specifically articulated rationale for why you have that expectation that's more than just "I prefer to be cautious" which is mostly a way of saying "I picked this number because I feel like it"

 

[petit_bateau]

I wouldn't call 75k/month for Shanghai "a dose of reality". Shareloft via Rob Maurer provided a credible report of September production of 82k, so you are predicting that the Q4 average is going to be 8.5% less than the rate already achieved a month ago, when the new production line upgrades had only recently been installed in August and there were likely still some kinks being worked out from the changes.

Your estimates imply a significant reversal of the long-term growth trend, which thus far has been interrupted only by government fiat and by planned production pauses for line upgrades. This would be the first time in the entire history of Shanghai in which a quarter without line stoppages had less output than the 3rd month of the prior quarter.

That prediction could be correct, but please present a specifically articulated rationale for why you have that expectation that's more than just "I prefer to be cautious" which is mostly a way of saying "I picked this number because I feel like it"

View attachment 865157

1. Per CPCA data July production from Shanghai was likely 30,257 m3 and 51,831 mY for a total 82,088.
2. If Shanghai can maintain 82k/m that would be 246/qtr.
3. We've seen Shanghai reduce in production QoQ before. The Q1-22 and Q2-22 quarters were slowdowns.
4. Same in Fremont, for example Q3-21, Q1-22
5. There is always a specific reason for it. Covid, line changes, supply chain shortages, major holidays and plant shutdowns, etc.
6. It is the same in every operation I've worked in / run, and I've run a few. There are always individual reasons, but overall it always happens.
7. Tesla themselves release very little structured data regarding the build-up down) of a plant's technical capability from the line/product level then adding in the normal stuff and finally coming out with what I would term an IPSC, "integrated production system capacity". In fact Tesla don't even release quarterly numbers broken down to plant level at all. Which leaves us trying to back snippets of data into large gaps that they cannot possibly fill.
8. For this reason I am cautious until I see repeated performance at any given level.
9. And there are far larger error terms on the table than Shanghai at the moment.

I don't expect that will make you happy, but that's how I see it.

Any thoughts on whether Fremont achieved the production increase I suggested ? or is it Berlin/Austin that managed that ? Nobody seems to know, or at least nobody who knows is saying anything, and I've not seen anyone other than myself draw attention to this puzzle. If we cannot figure out what went on in Q3 then forecasting Q4 is going to be a bit tricky even if we were to call Shanghai correctly.

 

[ammulder]

Any thoughts on whether Fremont achieved the production increase I suggested ? or is it Berlin/Austin that managed that ? Nobody seems to know, or at least nobody who knows is saying anything, and I've not seen anyone other than myself draw attention to this puzzle. If we cannot figure out what went on in Q3 then forecasting Q4 is going to be a bit tricky even if we were to call Shanghai correctly.

Its possible the Berlin Q3 estimate of 15k is low. I don’t have hard numbers but I feel like they hit 1000/wk early in Q3 and 2000/week just now (either end of Q3 or early Q4). If that wasn’t dramatically back-loaded then the total could be more like 18k? But again, just a possibility. Has anyone been tracking the highest Berlin VINs seen or anything like that?

 

[BrownOuttaSpec]

"Analyst" continue to revise their estimate downward today the EPS was updated. It's down another $0.01 since Wednesday, now EPS is estimated at $1.00 on yahoo. I would assume we will drop another 1-2¢ before the actual earnings are announced with the current rate we've seen since the P&D was announced. Get you ready for Oct 19th

View attachment 863691View attachment 863694

I'm wondering if Analyst are that naive or they are setting up the market for a recovery by settings expectations too low on purpose.

Looks like we did it, right at the buzzer, continued to slide down and now have an EPS is 0.99

Don't forget to put in your guess if you want to play along at home (Thanks Gigapress! link):
TeslaMotorsClub Investor Casual Quarterly Estimates Compilation

 

[Yuha]

1. Per CPCA data July production from Shanghai was likely 30,257 m3 and 51,831 mY for a total 82,088.
2. If Shanghai can maintain 82k/m that would be 246/qtr.
3. We've seen Shanghai reduce in production QoQ before. The Q1-22 and Q2-22 quarters were slowdowns.
4. Same in Fremont, for example Q3-21, Q1-22
5. There is always a specific reason for it. Covid, line changes, supply chain shortages, major holidays and plant shutdowns, etc.
6. It is the same in every operation I've worked in / run, and I've run a few. There are always individual reasons, but overall it always happens.
7. Tesla themselves release very little structured data regarding the build-up down) of a plant's technical capability from the line/product level then adding in the normal stuff and finally coming out with what I would term an IPSC, "integrated production system capacity". In fact Tesla don't even release quarterly numbers broken down to plant level at all. Which leaves us trying to back snippets of data into large gaps that they cannot possibly fill.
8. For this reason I am cautious until I see repeated performance at any given level.
9. And there are far larger error terms on the table than Shanghai at the moment.

I don't expect that will make you happy, but that's how I see it.

Any thoughts on whether Fremont achieved the production increase I suggested ? or is it Berlin/Austin that managed that ? Nobody seems to know, or at least nobody who knows is saying anything, and I've not seen anyone other than myself draw attention to this puzzle. If we cannot figure out what went on in Q3 then forecasting Q4 is going to be a bit tricky even if we were to call Shanghai correctly.

One problem with your math: September is a 30 day month and there were production interruptions. September 10-12 was a three day holiday weekend called Mid Autumn Festival and according to Wu Wa Tesla was operating with reduced shifts. On September 14 a typhoon hit Shanghai and Tesla factory was closed during it. Assuming they lost two days of production because of those interruptions, that leaves 28 days or exactly 4 weeks of production. That's 20,522 per week.

Q4 is 13 weeks and one day. If we assume the same production rate and no interruptions, we would expect production of about 270k. Of course it is unreasonable to assume no interruptions, so something like 250k seems reasonable, or maybe 240k if you want to be extra conservative.