Near-future quarterly financial projections

[DarkandStormy]

https://twitter.com/x/status/1546133019563638784

This isn't the same thing since it's taken from July of 2022, but ~6 months ago, FY22 non-GAAP EPS estimates were around $12. Tesla is expected to report ~$4 after the close tomorrow for FY22 (and at one point, 4Q EPS estimates were at $4). You can obviously see FY23 EPS have fallen dramatically. So perhaps the analysts have over-adjusted after missing 2022 by such a large margin?

EDIT - disregard, I cannot keep all the stock splits straight lol

 

[The Accountant]

My Q4 Estimate shows record Revenues, Gross Profit $, Net Income and EPS.
Versus prior year, operating expenses were well managed considering the inflationary environment.
GAAP EPS of $1.12 would be 65% above last year's number of $0.68
Non-GAAP EPS of $1.23 would be 46% above last year.

 

[The Accountant]

Here is my Q4 estimate along side the prior quarters of 2022:

 

[The Accountant]

Versus Q3, I expect Profit Margins % for auto to be flat at around 26.6% (excludes Reg Credits, Semi and Leases)
. . . . but Profit Margin dollars down $502 per car.

Q4 ASP is down $2k from Q3 driven by the following items:

Q4 Cost per Vehicle is down vs Q3 by $1.6k driven by the following items:

 

[The Accountant]

I have forecasted margins as follows:

 

[Chunky Jr.]

will there be a bitcoin impairment in Q4?

 

[The Accountant]

will there be a bitcoin impairment in Q4?

Yes - I have it at $38m which is only $0.01 impact on eps.
I wonder if they surprise us with a sale of the final stake.

 

[MP3Mike]

will there be a bitcoin impairment in Q4?

I haven't watched it, but that is probably answered in this video:

 

[Doggydogworld]

Looking at that Wall St average is only dialling in a rerun of Q4-2022 four times over. Why do they bother doing this ? They obviously know how to do it better. Is this laziness or conspiracy ?

Q4 is seasonally strong. Annualizing Q4 and adding 10% has historically worked fairly well for Tesla, except when they launched Model 3. With no high volume new product this year, that'd be my starting point for 2023. So 1.8m deliveries and 105b-ish revenue. Maybe 1.9m and 110b if Tesla qualifies for the full 7500 in the US after March.

 

[petit_bateau]

Q4 is seasonally strong. Annualizing Q4 and adding 10% has historically worked fairly well for Tesla, except when they launched Model 3. With no high volume new product this year, that'd be my starting point for 2023. So 1.8m deliveries and 105b-ish revenue. Maybe 1.9m and 110b if Tesla qualifies for the full 7500 in the US after March.

Yes, that might have been a fair approach in the past where there was a fairly close correlation between [new product] and [new production line].

However we are moving into a 2023 where the [existing] product is relatively constant (3,Y) and yet the [new production lines] ought to be ramping in a fairly material manner.

---

I do get the feeling that Tesla Investor Relations primary role is to get Wall St consensus earnings into the right landing zone by the last two weeks of each quarter.

 

[heltok]

Here is my Q4 estimate along side the prior quarters of 2022:

View attachment 899227

Great work. A quick idea. These figures are what Tesla has to report so they are useful to understand how low IQ wallstreet will react. But I care less about that and more about the fundamentals behind the numbers and what they say about the future. What if we ignored the waves and scaled deliveries to be equal to production in every quarter? How would these numbers look like? Is it hard to do in your spreadsheets?

 

[Artful Dodger]

What if we ignored the waves and scaled deliveries to be equal to production in every quarter?

What if we ignored the noise, and estimated delivery volume as 10-days less than production rate?

I think having 10-days of inventory at the end of each quarter is a healthy level.

 

[heltok]

What if we ignored the noise, and estimated delivery volume as 10-days less than production rate?

I think having 10-days of inventory at the end of each quarter is a healthy level.

Sure that makes sense. My point was just that there has been so much noise with pandemics, shipping, ending the wave etc that makes the data noisier and harder to do inference. Any way of just removing this is fine. Your method would also remove the noise, and maybe feels better for some people. I don't really care.

 

[JohnnyEnglish]

What if we ignored the noise, and estimated delivery volume as 10-days less than production rate?

I think having 10-days of inventory at the end of each quarter is a healthy level.

????? I think your logic is askew here.

If you have delivery volume as 10 days less than production rate each quarter then the inventory will just keep building up (roughly 40 days each year).

If you have 10 days of inventory at the end of each quarter then a 50% annual compound growth in production leads to deliveries being ~ 99% of production so given our modelling accuracy it would be equivalent to having deliveries = production.

 

[petit_bateau]

????? I think your logic is askew here.

If you have delivery volume as 10 days less than production rate each quarter then the inventory will just keep building up (roughly 40 days each year).

If you have 10 days of inventory at the end of each quarter then a 50% annual compound growth in production leads to deliveries being over 99% of production so given our modelling accuracy it would be equivalent to having deliveries = production.

I agree with your logic.

On a related matter have we actually figured out the Dec China production data from Shanghai ? I've not seen it, only the local sales data.

 

[JohnnyEnglish]

I agree with your logic.

On a related matter have we actually figured out the Dec China production data from Shanghai ? I've not seen it, only the local sales data.

From Rob Maurer production figure and his estimated split 3/Y. MIC December production 58,781. Model 3 21,500, Model Y 37,300

 

[The Accountant]

Great work. A quick idea. These figures are what Tesla has to report so they are useful to understand how low IQ wallstreet will react. But I care less about that and more about the fundamentals behind the numbers and what they say about the future. What if we ignored the waves and scaled deliveries to be equal to production in every quarter? How would these numbers look like? Is it hard to do in your spreadsheets?

If I insert Production numbers in place of deliveries in my model for Q3 and Q4, non-GAAP eps would have changed:

Q3 from $1.05 to $1.15
Q4 from $1.23 to $1.36

 

[petit_bateau]

From Rob Maurer production figure and his estimated split 3/Y. MIC December production 58,781. Model 3 21,500, Model Y 37,300

Thanks.
====
Q1. Where did it stand in Nov ? It was over 100k sales, but that might have been 90k from production and 10k from stock - do we know ?
===
For info this is my best understanding of the respective ramp rates to date

===
Q2. What is our best understanding of when and how many CT will launch ?

 

[petit_bateau]

I'm setting up my 2024 quarterly tables. I run my spreadsheets with two years ahead in quarters, and thereafter in annuals.

Grumble : The Semis need to be included in the Tesla quarterly P&D otherwise it screws up the vehicle counts. They are vehicles, fight for their right to be counted !

As we all know it is all about cell supply. It is cell supply that is the most significant constraint on staying on the ramp rate required to achieve the 20m/yr by 2030 target. That said there are also some looming problems if they don't start committing to additional lines and/or additional models and/or additional factory sites fairly soon.

For now I am assuming that Shanghai, Berlin, and Austin keep on adding & ramping 3/Y lines. Also I assume a pacy CT ramp similar to the early days of the 3.

However that still leaves a problem. My reckoning is that 2024 needs to be somewhere in the range 3.5m - 3.8m vehicles per year. The 3.5m corresponds to a 50% ramp rate per year, but that is actually at the low end of the range for the middle of the required S-curve, and so 3.8m is a better mid-decade target. By my calcs they are behind with CT production start (and behind with the 4680 ramp) and that this in turn is impacting and delaying two subsequent model platforms, namely the 2/Z and the van. If I slip their first production a year into 2024 rather than 2023, the result is a shortfall in 2024 volumes that is noticeable.

This suggests that the next gen platform announcements are going to be in this area.

There is a corresponding issue with factory space. I sense another factory announcement coming soon (I don't mean the Semi in Nevada), plus we have rumours of delays in announcing a Shanghai expansion.

Has anyone any further / better insight ?

Otherwise I think Tesla will not make it to the 20m/yr by 2030 target.

 

[DarkandStormy]

I'm setting up my 2024 quarterly tables. I run my spreadsheets with two years ahead in quarters, and thereafter in annuals.

Grumble : The Semis need to be included in the Tesla quarterly P&D otherwise it screws up the vehicle counts. They are vehicles, fight for their right to be counted !

As we all know it is all about cell supply. It is cell supply that is the most significant constraint on staying on the ramp rate required to achieve the 20m/yr by 2030 target. That said there are also some looming problems if they don't start committing to additional lines and/or additional models and/or additional factory sites fairly soon.

For now I am assuming that Shanghai, Berlin, and Austin keep on adding & ramping 3/Y lines. Also I assume a pacy CT ramp similar to the early days of the 3.

However that still leaves a problem. My reckoning is that 2024 needs to be somewhere in the range 3.5m - 3.8m vehicles per year. The 3.5m corresponds to a 50% ramp rate per year, but that is actually at the low end of the range for the middle of the required S-curve, and so 3.8m is a better mid-decade target. By my calcs they are behind with CT production start (and behind with the 4680 ramp) and that this in turn is impacting and delaying two subsequent model platforms, namely the 2/Z and the van. If I slip their first production a year into 2024 rather than 2023, the result is a shortfall in 2024 volumes that is noticeable.

This suggests that the next gen platform announcements are going to be in this area.

There is a corresponding issue with factory space. I sense another factory announcement coming soon (I don't mean the Semi in Nevada), plus we have rumours of delays in announcing a Shanghai expansion.

Has anyone any further / better insight ?

Otherwise I think Tesla will not make it to the 20m/yr by 2030 target.

View attachment 899691

Fwiw, for FY23 @Troy is at ~1.836m deliveries, TSLA IR compiled average is 1.849m, Gary Black is at 1.9m, and WS average (per Bloomberg, as of two days ago) is 1.85m. Will be interesting to see how much those change after the EC today. Is WS 23% low?

Not sure how likely we are to see 2/Z and/or Van vehicles at all in FY2024. The last couple of iterations - Semi & CT - have taken 3-4 years (or longer) from product unveil to first deliveries. The only one that went quickly was MY, but that was mostly because of how many components it shares with the Model 3, I think.

You asked about CT ramping in your previous post - I think this is a big variable. It's a new method of manufacturing so we don't really have a historical basis from Tesla to use. It could go quickly or there could be a lot of initial bugs that slow things down. I would say this ramp is going to be very hard to predict for that reason.

Barring updated guidance from the Tesla team, I think it's just a lot of shots in the dark right now -> perhaps we'll hear more about their plans later today or on March 1st and get a clearer picture.