Near-future quarterly financial projections

[dc_h]

Well, that's about 35K cars, right? (Rough math: 450,000/13) Maybe 40K if they hit 495K production and it's back-loaded? That's not so incredibly far from Troy's numbers, but it's still a higher discrepancy than I'd expect. By that logic, they have the last week of production from Q3 to deliver in Q4, so it would really only be the difference between a week of end-of-Q4 production and a week of end-of-Q3 production, which would be 10K or under, right? And we already know there were 20K+ produced but not delivered at the end of Q3. Seems like a pretty bad case would be they can't deliver 40K from end-of-Q4 production, but they do deliver 20K of Q3 production, leaving a ~20K discrepancy.

If this *was* all caused by China demand unexpectedly falling slightly under upgraded production capacity -- that is, not enough deliveries to perform really close to Shanghai to soak up its end-of-quarter production -- AND if that recurs in Q4 (neither of which is a given) -- then I'd expect Tesla to better manage that in Q4. Geographically, they have demand all over the place, and they can manage production/shipping accordingly. Maybe they put more on trains in early December so they can deliver more really near the factory in late December, or whatever.

I think Troy and others are predicting that at least 40,000 will be in transit due to limitations on Shanghai shipping capacity. They need to smooth out RORO deliveries at maintain about 40,000 to 45,000 every month, they can't physically front load the quarters anymore, since the plant has more capacity than the harbor. If they smooth out delivery centers, that could be another 15,000 or more. So basically adding 30,000 to 40,000 on top of the 20,000 in Q3. Gordon Johnson will jump on it, but with Tesla at this size and share buybacks on the table, Wall Street might be more forgiving about targets (I know, it's never happened before, but maybe this time). Bottom line, if they hit the stretch 495k, 455-475k would be max deliveries. 425-450 seems more likely.

 

[thx1139]

I think Troy and others are predicting that at least 40,000 will be in transit due to limitations on Shanghai shipping capacity. They need to smooth out RORO deliveries at maintain about 40,000 to 45,000 every month, they can't physically front load the quarters anymore, since the plant has more capacity than the harbor. If they smooth out delivery centers, that could be another 15,000 or more. So basically adding 30,000 to 40,000 on top of the 20,000 in Q3. Gordon Johnson will jump on it, but with Tesla at this size and share buybacks on the table, Wall Street might be more forgiving about targets (I know, it's never happened before, but maybe this time). Bottom line, if they hit the stretch 495k, 455-475k would be max deliveries. 425-450 seems more likely.

Troy calculation is far more than 40K in transit. At end of Q3 inventory was like 35K based on Rob Maurer calculations. Troy has 400K deliveries and 454K production. Based on that the amount in transit would be like 90K.

 

[JohnnyEnglish]

Do we have any info on how many shifts Tesla are operating at Giga Texas? Elon's comments about them being close to achieving 2k/week seems too high for a single shift, much more likely to be 2 shifts.

 

[petit_bateau]

Well I was rather high on my Q3 estimate at 14% above the actual.

ITEM -------ESTIMATE (post P&D) --- ACTUAL
Revenue = $23.204 bn ===> $21.454
Auto GM% = 29.0% ===> 27.9%
EPS (GAAP) = $1.09 ===> $0.95 (so I was 14% above the actual)

Looking backwards it seems everyone is really unsure what the production was out of Fremont, Berlin, Austin. Looking forwards I've listened to the various views on likely Q4 production out of Shanghai as a result I've now taken a best of Q3 monthly rate and applied a 95% modifier to it. The Austin and Berlin ramps I've dialled in are based off the limited info we have after applying some moderation of optimism. The GM% I held at 28% as I really don't know how the balance of price-changes and input costs and ramp costs will play out. I assume the 8d (30k cars) inventory is held constant exiting Q4 as I'm sure Tesla will try very hard not to let it grow to 40k cars, and therefore it is a wash going forwards into Q1.

For full-year 2022 this suggest 1.39m vs 0.93m production from 2021, i.e. 49% yoy increase. Tantalisingly close, we'll see.

For solar it seems to me that unless they make a significant product breakthrough (unlikely imho) or a radical business strategy change (easy, but sadly unlikely) then a volume and profitability change is unlikely and so a plateau of 100MW/qtr seems most probable. For storage it seems that Lathrop is getting onto the ramp and also that power semiconductor shortages are easing, and so I am more optimistic.

With much the same for services & other, and overheads, this yields the following financials

in turn giving

There is then a real tension in the TSLA share market between participants who use PE, or PEG, or NPV, and whether forwards or trailing as they come to very different share price outcomes, most especially in the second half of next year. As an example the below shows what happens if the trailing PE settles at ~54, but note that puts trailing PEG under pressure in both directions through the course of the year as it would vary from 1.0 to 1.8. Hence there is considerable scope for a very erratic shareprice as the market tries to satisfy the contradictory - but equally valid - views of many different TSLA participants. And that is even with no macro sentiment input. I guess the wild ride is not over.

The annual version below gives some end-year fair value share prices driven by a NPV calculation with a 10% discount rate for comparison.

 

[The Accountant]

My Q3 Tesla estimate vs actual results.
My biggest miss was on Auto Revenue where my Average Selling Price per car was too high.
Mix skewed more toward lower trim models than I modeled.
I got lucky on COGS. If I was to adjust the Auto Revenue for the lower trim models, than my COGS would have been lower.
So I missed on skew mix but also on auto margins. Berlin and Austin had slightly lower margins than I modeled.

 

[Artful Dodger]

My Q3 Tesla estimate vs actual results.
My biggest miss was on Auto Revenue where my Average Selling Price per car was too high.

This discussion seems to 'bury the lede'. Can you tease out the affect of ForEx on Auto Revenue, and where a better estimate on ForEx changes would have left your Q3 estimate?

To me, the remaining large uncertainty is the product mix from Shanghai (SR+ vs LR), and how that affects ASPs. Not sure if that can be deduced even in hindsight. Any thoughts?

Thanks!

 

[XHabjab]

[...]
Looking backwards it seems everyone is really unsure what the production was...

My Q3 Tesla estimate vs actual results.
My biggest miss was on Auto Revenue where my Average Selling Price per car was too high...

No charts from me. I'm just a simple country boy...

Taking a cursory look at your numbers, your spreadsheet numbers seem to be based on production numbers. Shouldn't all the ASP and revenue etc. be based on deliveries? After all, no money changes hands until a delivery happens. I see this as an issue starting with this quarter because Elon has said that they will be unrolling the wave, and will reduce or eliminate the on-again off-again shipments with RORO ships (at least), allegedly because they can't get availability on the Tesla quarterly schedule. What this means is that more cars will be seen on the books as inventory and a smaller percentage of produced cars delivered at the all important end of quarter. Thus the larger than seen before gap between Q3 production and deliveries. I see this more importantly as a D/P ratio rather than a P-D gap size.

To contribute to this, there were some rumours that the Shanghai factory was directed to decrease production for a few days at the end of the quarter. If this was true, my suspicion is that this sand bagging would bring quarterly production and deliveries closer together (and compress the D/P ratio in the direction of 100%).

I don't think the wave is completely unrolled and this quarter's D/P ratio does not represent the long term value. I would imagine that availability of car carriers and trains for example might see an unroll as well. Also, the wave was mostly unrolled at the Shanghai factory. My simplistic prediction says that we will see unsold inventory increase for Q4 as well, showing a smaller ratio for Q4 and beyond (tempered each time by unsold cars from previous quarter sold/delivered in 'this' quarter). I think we can only determine the long term value in retrospect, sometime in the future.

 

[The Accountant]

No charts from me. I'm just a simple country boy...

Taking a cursory look at your numbers, your spreadsheet numbers seem to be based on production numbers. Shouldn't all the ASP and revenue etc. be based on deliveries? After all, no money changes hands until a delivery happens. I see this as an issue starting with this quarter because Elon has said that they will be unrolling the wave, and will reduce or eliminate the on-again off-again shipments with RORO ships (at least), allegedly because they can't get availability on the Tesla quarterly schedule. What this means is that more cars will be seen on the books as inventory and a smaller percentage of produced cars delivered at the all important end of quarter. Thus the larger than seen before gap between Q3 production and deliveries. I see this more importantly as a D/P ratio rather than a P-D gap size.

To contribute to this, there were some rumours that the Shanghai factory was directed to decrease production for a few days at the end of the quarter. If this was true, my suspicion is that this sand bagging would bring quarterly production and deliveries closer together (and compress the D/P ratio in the direction of 100%).

I don't think the wave is completely unrolled and this quarter's D/P ratio does not represent the long term value. I would imagine that availability of car carriers and trains for example might see an unroll as well. Also, the wave was mostly unrolled at the Shanghai factory. My simplistic prediction says that we will see unsold inventory increase for Q4 as well, showing a smaller ratio for Q4 and beyond (tempered each time by unsold cars from previous quarter sold/delivered in 'this' quarter). I think we can only determine the long term value in retrospect, sometime in the future.

Not sure if your first comment was meant for me. My earnings forecasts are based on deliveries.

regarding the eave, I agree with you that the flattening of the wave will be done gradually over the next several quarters and not completed in Q4. I think we may see a discrepancy between production and deliveries of about 15k in Q4 +/- 5k.

 

[petit_bateau]

No charts from me. I'm just a simple country boy...

Taking a cursory look at your numbers, your spreadsheet numbers seem to be based on production numbers. Shouldn't all the ASP and revenue etc. be based on deliveries? After all, no money changes hands until a delivery happens. I see this as an issue starting with this quarter because Elon has said that they will be unrolling the wave, and will reduce or eliminate the on-again off-again shipments with RORO ships (at least), allegedly because they can't get availability on the Tesla quarterly schedule. What this means is that more cars will be seen on the books as inventory and a smaller percentage of produced cars delivered at the all important end of quarter. Thus the larger than seen before gap between Q3 production and deliveries. I see this more importantly as a D/P ratio rather than a P-D gap size.

To contribute to this, there were some rumours that the Shanghai factory was directed to decrease production for a few days at the end of the quarter. If this was true, my suspicion is that this sand bagging would bring quarterly production and deliveries closer together (and compress the D/P ratio in the direction of 100%).

I don't think the wave is completely unrolled and this quarter's D/P ratio does not represent the long term value. I would imagine that availability of car carriers and trains for example might see an unroll as well. Also, the wave was mostly unrolled at the Shanghai factory. My simplistic prediction says that we will see unsold inventory increase for Q4 as well, showing a smaller ratio for Q4 and beyond (tempered each time by unsold cars from previous quarter sold/delivered in 'this' quarter). I think we can only determine the long term value in retrospect, sometime in the future.

I calculate ASP and revenue based on deliveries. But the table I showed for production includes the model & factory mix and I showed it because it is more interesting and we have a better feel for the input data. I don't have a corresponding table of deliveries per factory and model as it would not add any value and be horrifyingly complex with far too many unknowns. So I just have a single forecast delivered number.

Getting from production to deliveries is easy when stock-in-transit is constant (either in % prod terms or absolute #units terms, albeit a different calc either way). However again the issue arises that we have no effective forecasting information on what that stock-in-transit number will be. At present it is 8d which corresponds to 30k vehicles. I have no reason to expect them to hold it at either level percentage 8d (which would be 38k in the coming quarter) or at 30k (level, absolute) or at any other particular number. Like you I don't think they have unwound the wave, indeed my direct personal observation was a wave of car transporter trucks carrying Teslas over the Brenner pass from the Slovenia port in the last days of the quarter, so quite the reverse - they were trying very hard to slosh the wave into the bucket.

Directionally I think their short-term preference is to maximise EOY delivery provided that they do not overpay for logistics as I think that although they know they should unwind the wave, they would rather be chaste in the future than now. Please make me pure, but not yet. It is only when external factors have imposed a constraint that the wave gets unwound (oops, no car carrier ships). Therefore my expectation is that they will likely hold the inter-continental wave steady (i.e. fixed # car carriers ships in transit at qtr end), but seek to manage intra-continental wave better to minimise EOQ stock. So for the time being I chose to hold EOQ stock constant at 30k which implies a progressively declining % stock (i.e. 8d >> 7d >> 6d) which can hold true if they learn fiendishly good logistics. I'll be watching that as it may be that in practice it goes in the other direction and if so I will switch my equation to 8d >> 9d >> 10d etc.

That is an interesting rumour you had out of Shanghai. There is certainly too little firm data to close the equation. Your info (if true) also supports my conclusion that Fremont is most likely the source of the extra units which means that Fremont is working even better. Certainly they came from somewhere. It is fascinating that nobody seems to know where they came from, at least nobody who can talk.

====
By the way I don't directly account for any forex impacts in my model. My view is that Tesla is fairly well balanced intrinsically (internally) and so ordinarily that is not a great deal. My fag packet calc was that the forex impact in Q3 was about half the size of the GM% impact, so occasionally my view gets caught out.

====
On the matter of the vehicles GM% impact they are essentially honouring car prices fixed up to a year ago, using material prices that were fixed as recently as ??? months ??? ago. Therefore there was (in retrospect) some GM% compression as a wave of raw materials price increases washed through the system. Perhaps now the raw material prices are receding (i.e. costs stable or decreasing) and selling prices that were raised back-a-ways will come roaring in. If so the 28% GM could rise higher, maybe >> 32% or whatever. But again we have no firm data and so I held my future forecast constant at GM=28% (I had been at 30% which proved too high in Q3). I noted that averaging the last 4 qtrs of GM% would also give a good match to the actual Q3, but had no good reason to choose either way.

Note there are continuing ramps to support on Y production, plus also now Semi, and they impact GM% adversely.

 

[adiggs]

I assume the 8d (30k cars) inventory is held constant exiting Q4 as I'm sure Tesla will try very hard not to let it grow to 40k cars, and therefore it is a wash going forwards into Q1.

I have nothing other than my gut on this one - my working assumption is that end of quarter inventory will be 10% of unit deliveries (EDIT: production!!), as a mechanism for accounting for unwinding and then performing consistently with the wave being unwound. This assumes that some fraction of recent production is undelivered, as well as all inventory entering the quarter has been sold.

I think some sort of relationship like this will need to be found - I expect that having the wave unwound will translate into a steadily increasing quarter to quarter inventory as long as production is steadily increasing quarter to quarter.

 

[JohnnyEnglish]

@The Accountant , @petit_bateau Are either of you including a release of deferred revenue associated with FSD going to wide release in North America? I have guessed that this might be around $500M(if it happens) in my model but would appreciate your views on this.

 

[petit_bateau]

I have nothing other than my gut on this one - my working assumption is that end of quarter inventory will be 10% of unit deliveries (EDIT: production!!), as a mechanism for accounting for unwinding and then performing consistently with the wave being unwound. This assumes that some fraction of recent production is undelivered, as well as all inventory entering the quarter has been sold.

I think some sort of relationship like this will need to be found - I expect that having the wave unwound will translate into a steadily increasing quarter to quarter inventory as long as production is steadily increasing quarter to quarter.

365/4 = 91 days per quarter
8/91 = 8.8% where 8-days is the end Q3 inventory, corresponding to about 30k
... but very crudely it is nigh-on 1-day = 1%

My suspicion is that they'll do their best to deliver as many cars as possible in Q4, which will tend to exacerbate the wave (i.e. send inventory below 8d) as it is human nature to try. However given the general constraint of not overpaying for delivery logistics, and the specific constraint of there being a finite number of car-carrying ships, I think they'll move towards the waves happening mostly within continents than between continents. Clearly this is a complex thing as all sorts of puzzles have to be solved in an optimal manner - e.g. Taiwan can't be supplied from Shanghai, or Berlin does not make model 3, or S/X only come from Fremont. We watch with interest.

@The Accountant , @petit_bateau Are either of you including a release of deferred revenue associated with FSD going to wide release in North America? I have guessed that this might be around $500M(if it happens) in my model but would appreciate your views on this.

I have no opinion on this and I don't include it. It is not because I am not hopeful about FSD (I am, but on a different timescale and with far wider implications). It is simply because there is no observable evidence that allows me to make a time-based modelling prediction. So it only shows up in my model as being one of the contributors to an unusually high GM%.

A related point I'd like to make is that many folk in US embed country-specific bias in their views due to over exposure to US and under exposure to Rest Of World. Sometimes that works well, for example observing Teslafornia adoption up close. Other times it is a blind-sider, for example thinkin the whole world wants black solar shingle roofs when really this is a unproductive dead-end for a looong time. It is possible that FSD is one of these cases as it simply isn't allowed in Europe or China as regulations currenty stand.

 

[adiggs]

365/4 = 91 days per quarter
8/91 = 8.8% where 8-days is the end Q3 inventory, corresponding to about 30k
... but very crudely it is nigh-on 1-day = 1%

My suspicion is that they'll do their best to deliver as many cars as possible in Q4, which will tend to exacerbate the wave (i.e. send inventory below 8d) as it is human nature to try. However given the general constraint of not overpaying for delivery logistics, and the specific constraint of there being a finite number of car-carrying ships, I think they'll move towards the waves happening mostly within continents than between continents. Clearly this is a complex thing as all sorts of puzzles have to be solved in an optimal manner - e.g. Taiwan can't be supplied from Shanghai, or Berlin does not make model 3, or S/X only come from Fremont. We watch with interest.


I have no opinion on this and I don't include it. It is not because I am not hopeful about FSD (I am, but on a different timescale and with far wider implications). It is simply because there is no observable evidence that allows me to make a time-based modelling prediction.

Part of that thumb-to-the-wind 10% is an expectation that manufacturing will continue to ramp throughout the quarter (and its simple). So more back loaded production than front loaded. Part of it is an observation I have that downtime for line optimization / expansion tends to be more front half of the quarter than back half.

The real point, whatever % that we use - I see the ending units each quarter as being a function of production in that quarter, and effectively independent of whatever the inventory was entering the quarter. This is really the idea that I was trying to add - I think its erroneous to forecast ending inventory / units as some # plus whatever ending inventory was for the previous quarter.


You're probably right about what they'll do with the wave. I continue to really, really hope that you're wrong. Heck - I'd even go so far as to say that I would like to see Tesla stop publishing a deliveries number. Just publish the production number at the start of the quarter; try to shift the conversation away from deliveries. If investors and media can be shifted to production as the focus point, then the wave can be easily erased in the back ground, not just from the financials but from any hint of a business process.

 

[petit_bateau]

Part of that thumb-to-the-wind 10% is an expectation that manufacturing will continue to ramp throughout the quarter (and its simple). So more back loaded production than front loaded. Part of it is an observation I have that downtime for line optimization / expansion tends to be more front half of the quarter than back half.

The real point, whatever % that we use - I see the ending units each quarter as being a function of production in that quarter, and effectively independent of whatever the inventory was entering the quarter. This is really the idea that I was trying to add - I think its erroneous to forecast ending inventory / units as some # plus whatever ending inventory was for the previous quarter.


You're probably right about what they'll do with the wave. I continue to really, really hope that you're wrong. Heck - I'd even go so far as to say that I would like to see Tesla stop publishing a deliveries number. Just publish the production number at the start of the quarter; try to shift the conversation away from deliveries. If investors and media can be shifted to production as the focus point, then the wave can be easily erased in the back ground, not just from the financials but from any hint of a business process.

Yeah I know what you mean. It is the normal internal Mexican standoff of logistics, production, sales, and finance. So for now I have kept my model simple and assumed a constant 30k cars. I'm sure I'm wrong, but not sure which way

 

[petit_bateau]

Berlin VIN analysis in this post by @hobbes might indicate that Berlin is the source of the extra production, not Fremont.

Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

Elon said his last sale was insurance in case he needed more for the Twitter purchase and that he wasn't planning to sell more. Why are so many people here convinced he is going to sell more? I can't speak for anyone else, but here are my thoughts. When he mentioned his last sale was...

teslamotorsclub.com

VIN analysis can mislead, but highest observed VIN is 36304 on Oct 24 in Norway. By comparison my guesstimated end Q3 Berlin cumulative production was 23,280. VIN analysis suggests current weekly rate is 2912/week, and a delay of about a week in getting to Norway. So maybe just over 2-weeks of production to correct for since end Q3 after allowing for transit lag, call that 6,000. So 36,304-6,000 = 30,000 vs the 23,000 I'd guesstimated, so perhaps Berlin was 7k ahead of where i/we thought it was. It is possible that Berlin is the missing source ..... but to what extent are there gaps in the VIN sequence and what are the other error terms ?

By comparison I had to add 7k to Fremont to close the production data gaps for Q3. So this does seem to all match up.

Hmmmmmmmmm..............

 

[JohnnyEnglish]

...

A related point I'd like to make is that many folk in US embed country-specific bias in their views due to over exposure to US and under exposure to Rest Of World. ...

I am in the UK, not the US

 

[The Accountant]

@The Accountant , @petit_bateau Are either of you including a release of deferred revenue associated with FSD going to wide release in North America? I have guessed that this might be around $500M(if it happens) in my model but would appreciate your views on this.

I don't have FSD Deferred Revenue release in my Q4 numbers. If we see a NA wide release by year end, I will include it for Q4.
I have not looked at the number recently but I seem to remember I was looking at a number higher than $500m . . more like $800m.
I will dig into this more when I complete my Q4 estimate.

 

[dc_h]

Just a note regarding the wave: They have to undo the wave in Shanghai due to shipping volume limitations, so that is going to add ~40,000 in transit above normal, global combined, in country numbers. They probably also have to level off Fremont shipping as that volume increases, but to a lessor extent. My understanding of the Shanghai shipping is that it can only handle about 42,000 cars per month. In theory they can backload shipments to Korea, Japan and Australia and front load Europe, but not sure it is worth the effort for Tesla logistics. Elon also stated he wants to reduce wait times, another nod to evening out shipping & leveling the wave.

 

[petit_bateau]

Just a note regarding the wave: They have to undo the wave in Shanghai due to shipping volume limitations, so that is going to add ~40,000 in transit above normal, global combined, in country numbers. They probably also have to level off Fremont shipping as that volume increases, but to a lessor extent. My understanding of the Shanghai shipping is that it can only handle about 42,000 cars per month. In theory they can backload shipments to Korea, Japan and Australia and front load Europe, but not sure it is worth the effort for Tesla logistics. Elon also stated he wants to reduce wait times, another nod to evening out shipping & leveling the wave.

Something I popped in today's Energy News is that now the Chine-Europe rail links have started accepting lithium batteries various manufacturers have started shipping BEVs (aka NEVs) by rail. As yet Tesla aren't doing this, but other auto mfg have started trials of the route. Transit times are about half that of ro-ro sea-freight so that is an additional factor. It tends to be the more inland factories that bias towards these lines, but quite a lot of stuff does originate out of Shangai. It is also possible this will be a route for LFP packs. A contrary factor might be a reluctance by Tesla to have anything significant in transit through Russia in the near future. I guess we just watch.

See post #345

Energy Sector News

Third way on MidCat, now BarMar / also 2.2GW Biscay link delayed https://www.montelnews.com/news/1360961/madrid-lisbon-paris-ditch-third-gas-link-project https://www.gem.wiki/Midi-Catalonia_Pipeline...

teslamotorsclub.com

and

China-Europe: new energy vehicles fever sees multiple services launched

Xi’an, Zhengzhou and Chengdu all launched new China-Europe rail services in October. Their common denominator? New energy vehicles (NEVs) occupy the train capacity and provide a glimpse of Silk Road cargo trends. The new cargo type is quickly taking over the supply chain since the Chinese...

www.railfreight.com

 

[JohnnyEnglish]

Had a mail from Tesla this morning saying the Model Y SR now available to order in the UK. Seems like they are expecting deliveries this quarter. At £6k less expensive than the LR AWD I would expect it to be very popular.

Nice Christmas presents for someone