Near-future quarterly financial projections

[mongo]

The annual limit of ~8 GWh 18,650 cell production in Japan comes from several independent sources:

• @hiroshiy mentioned it in this post. Hiroshi-san, are you able to share the source of your figure?

• Tesla made the following disclosure in the Q1'18 10-Q:
  • "We expect Model S and Model X deliveries to be approximately 100,000 in total
    in 2018, constrained by the total available supply of cells with the 18650 form factor used in those vehicles."

• The 8 GWh figure is a relatively straightforward estimate that comes from the owner-reported battery selection mix: about ~60% were 75D, ~40% 100D packs. If we apply this to the reported 99,475 Model S/X's delivered in 2018, we get: 99,475*0.60*75+99,475*0.40*100 = 8,455,375 kWh, or 8.4 GWh. Given that the survey is self-selecting and not statistically representative sample, I think it's probably a safe assumption that lower trim cars are under-reported vs. higher-trim cars - i.e. 75D sales were more than 60% and [P]100D sales were less than 40%. This would put the 2018 18,650 supply somewhere below 8.4 GWh. (I believe @ReflexFunds estimated a similar mix?)

• Finally, we also know that the supply limit is a capex constrained hard limit: this is how much the Japanese factories of Panasonic are able to manufacture per year. This was explained in the Q4'17 conference call by JB:
  • "Romit Jitendra Shah - Nomura Instinet
    
    Yes. Thank you. It sounds like from the letter that you could do more than 100,000 S and X in 2018, but you're constrained by the 18650s. And I'm just curious what would it take to see the 2170 cells in these vehicles?
    
    Elon Reeve Musk - Tesla, Inc.
    
    Yeah.
    
    Jeffrey B. Straubel - Tesla, Inc.
    
    Well, this is JB. It's something we've of course contemplated, but it's quite a large change to the architecture of the module and the battery pack overall. And while the 18650 supply is somewhat of a cap at about 100,000 units per year, even just a few months ago we didn't feel that expanding and making some long-term bets on expanding that supply with Panasonic in Japan was really the right risk. It's something we could consider, but right now we're pretty happy with that balance and it matches our other production capabilities and our other investments."
  • I.e. in 2017 Tesla contemplated whether to expand the 18,650 supply, and decided against it.

So I believe all of this is a pretty robust basis to believe that the 18,650 cell supply from Panasonic was somewhere in the 8.0-8.5 GWh range in 2018, and that after about ~10 years of factory optimizations and a mature industrial process it probably doesn't have much space for low capex supply expansion.

If Tesla really wanted to expand volume, and their supplier agreement allowed, Samsung previously provided 18650 cells to Tesla for evaluation before being sourced with 2170s for the Australian Energy project. So, if Samsung had capacity, Fremont could make more than 8.5 GWh of S/X.

One of many articles (but this one makes me wonder if TE can use 18650 and 2170)

Exclusive: Tesla Takes Delivery Of 1.74 Million Samsung Battery Cells

 

[Fact Checking]

If Tesla really wanted to expand volume, and their supplier agreement allowed, Samsung previously provided 18650 cells to Tesla for evaluation before being sourced with 2170s for the Australian Energy project. So, if Samsung had capacity, Fremont could make more than 8.5 GWh of S/X.

Samsung is unlikely to be able to offer the kind of bulk pricing Tesla very probably receives from Panasonic: Tesla's 18,650 contracts are probably long term and high volume, with guaranteed purchase quotas.

The Australian battery project was a showcase, at a net loss to Tesla.

 

[mongo]

Samsung is unlikely to be able to offer the kind of bulk pricing Tesla very probably receives from Panasonic: Tesla's 18,650 contracts are probably long term and high volume, with guaranteed purchase quotas.

The Australian battery project was a showcase, at a net loss to Tesla.

Could be, but Tesla was also airlifting Powerpacks to Australia (based on Elon's knowledge of the cost)

 

[winfield100]

Could be, but Tesla was also airlifting Powerpacks to Australia (based on Elon's knowledge of the cost)

South Australia gets 50 new grid-scale energy storage proposals
"......Tesla is also working on a giant “virtual power plant” that aims to have 50,000 homes with solar and storage linked. The first two stages aim to install 1,100 while the final stage is still subject to finance and a retail partner. AGL is also completing a separate scheme with 1,000 homes linked with solar and storage....."
"..Home Battery Scheme aims to install 40,000 batteries across the state.."
“The Marshall Government’s Grid Scale Storage Fund "
Australia, specifically SWA is leading
first a small i
project on tiny island'
then T'au
next Kaua'i
things going on quietly in Puerto Rico
Australia (a really bigger island)
next, Earth (Terra) a really big island
there is a rapid transition occurring everywhere

 

[Buckminster]

So, my understanding is that the excellent ReflexFunds' model is not as optimistic as it could be in at least 2 key areas:
1) S/X profit

So, in that case, isn't a $705m to $520m drop in profit unlikely?

Do we have any demand data yet?

2) M3 production + having 10k in delivery at end of quarter

BTW., assuming that Carsonight's GF1 production data continues to be as reliable as it was in the past, this suggests the following Q1 production rate:

• He reported more than 6k/week battery pack production sustained in the first ~7 weeks of Q1,
• recent production "approaching but not passing" 7k/week, which wording would suggest at least 6.7k/week,
• I think an average pack production rate of 6.4k-6.5k/week in Q1 so far is a conservative guess.
• This means ~45k Model 3 packs were already produced in Q1.
• If they keep 6.5k/week for the remaining 6 weeks (which is not a certainty: they might slow down at the end of Q1), that extrapolates to 82k Model 3's made in Q1, or +34% growth over Q4's 61.4k.

This is also supported by VIN allocations data: 98.5k VINs in Q1, which with the 85% estimate suggests a Q1 Model 3 production target of 83k units.

This is bullish: for example @ReflexFunds's Q1 estimates of tiny profits were based on Q1 production of 67k Model 3's. More Model 3's made will improve Q1 results - assuming they can be delivered. If "in-transit" vehicles inventory balloons over the guided "10k" range, that could be a drain on cash.

Also note: a +~35% jump in production could also generate cash, through payables expansion: with about $20k cash per extra unit made in the last 2 months of Q1, if there's ~15k more units then that would be about +$300m more cash.

So unless Tesla slows down production in the rest of Q1 dramatically, a 5k/week sustained rate of production seems secured, and 6k/week possible.

 

[generalenthu]

SEC Filing | Tesla, Inc.

10k

 

[brian45011]

The Australian battery project was a showcase, at a net loss to Tesla.

Very likely. The 10k shows 2018 Revenue for that project as $81.2 million and 2018 COGS as $72.5 million, for a gross margin of $8.7 million or 10.7%. However, the 4Q17 SH letter stated Air Freight (undisclosed amount) for that project was booked in 4Q17. After a reasonable allocation of SG&A (including Elon's several trips to OZ with visits with Amber on set while there), it's likely that project was right around or even slightly below break even.

But...as the MasterCard commercials use to say: the value of the favorable publicity was "PRICELESS."

 

[srius]

Samsung is unlikely to be able to offer the kind of bulk pricing Tesla very probably receives from Panasonic: Tesla's 18,650 contracts are probably long term and high volume, with guaranteed purchase quotas.

The Australian battery project was a showcase, at a net loss to Tesla.

Tesla should still source cells from Samsung/LG for Powerwall. Given the current demands and competitions, Tesla can raise price compensating higher cell cost to maintain the same margin. Powerwall production is still so limited even though the product was unveiled almost four years ago.

 

[schonelucht]

Can we estimate current Tesla cash position? Starting from last day of 2018 when they had $3,685 M in unrestricted cash.

What	Amount
Unrestricted cash year end	+ 3,685M​
Acounts payable paid	- 3,404M​
Acounts receivable received	+ 949M​
2 months SG&A	- 442M​
2 months R&D	- 237M​
2 months service revenue	+ 353M​
2 months service expense	- 445M​
2 months stock based compensation	- 125M​
15 000 M3 revenue	+ 900M​
7 000 S/X revenue	+ 700M​
4 000 trade ins	- 120M​
2 months of energy revenue	+ 247M​
Convertible	- 920M​
Unrestricted cash March 1st	1,141M​

Some assumptions

• Payment terms of 60 days both for receivables and payable, so everything paid&received. Accounts receivable remained high despite the quarter not ending on a week day. If the accounts payable has a large leasing partner in distress, this could put downwards pressure on cash received of several hunderds of millions.
• Assuming 20% do a trade in with a relative large body of tradeins high value cars (like off lease S/X) at $30K on average. Note sure how reliable this number is but could swing $100M either way
• Assumed all of the costs of goods for building cars/energy products in the first two months are covered either as an accounts payable item or through the asset back line. It's possible some suppliers aren't as generous and are in a position to enforce tougher terms and some payments were already due.
• Assumed most R&D and SG&A are costs that are directly payable like salary, leases and direct costs to run the organisation. But compensated for stock based salary.
• Assuming the capex ramp up for China has not yet started in a big way
• Assuming no increase in inventory for parts to improve service as earlier promised. Could swing a few tens of millions to the downside.

Are there big ticket items that I missed?

 

[schonelucht]

Accounts payable are paid back continuously, but while production is ongoing at the same level, this is offset by an equal amount of new payables generated.

Which is why I assume steady state production meaning accounts payable are paid, but no new cost of goods are accounted for production following the accounts payable.

Your numbers are equivalent to Tesla repaying $3.4bn payables and then inexplicably choosing to pay all suppliers immediately rather than using their agreed 60-90 days payable terms.

Incorrect. There is no line in my table for 'then inexplicably choosing to pay all suppliers immediately.'. Note that there is a line item for the full gross revenue value of 15k M3 and 7k S/X delivered so far with no corresponding cost of goods. There is also no cost of goods for the M3's and S/X's produced this quarter on top of those that have been delivered. That amount corresponds to the growth in accounts payable.

Please rebut or acknowledge that this was a fundamental misunderstanding of my method by you.

We can also do the exercise differently. Forget about accounts payable and accounts receivable and assume steady production. That would mean we would need to include in our table cost of goods for two months of production worth ($3.4B) as a minus on the cash position instead. To get to $2B you project, that would mean they somehow are able to stuff an additional $1,800M in the ABL and accounts payable? I am sorry, but that belies all belief.

 

[sandervl]

Are there big ticket items that I missed?

Your revenue for Model 3 & S+X seems too high, but the number of deliveries way too low. Assuming production is the same as in Q4/18 (unlikely), they should have produced around 60k cars already. 38k cars en route to customers seems a bit high.

InsideEV reports 12k Model 3 sales in the US for January & February.

Nevertheless it seems they're running a bit close to the edge this quarter, but this was predictable and acknowledged by Tesla.

 

[generalenthu]

Can we estimate current Tesla cash position? Starting from last day of 2018 when they had $3,685 M in unrestricted cash.

What	Amount
Unrestricted cash year end	+ 3,685M​
Acounts payable paid	- 3,404M​
Acounts receivable received	+ 949M​
2 months SG&A	- 442M​
2 months R&D	- 237M​
2 months service revenue	+ 353M​
2 months service expense	- 445M​
2 months stock based compensation	- 125M​
15 000 M3 revenue	+ 900M​
7 000 S/X revenue	+ 700M​
4 000 trade ins	- 120M​
2 months of energy revenue	+ 247M​
Convertible	- 920M​
Unrestricted cash March 1st	1,141M​

Some assumptions

• Payment terms of 60 days both for receivables and payable, so everything paid&received. Accounts receivable remained high despite the quarter not ending on a week day. If the accounts payable has a large leasing partner in distress, this could put downwards pressure on cash received of several hunderds of millions.
• Assuming 20% do a trade in with a relative large body of tradeins high value cars (like off lease S/X) at $30K on average. Note sure how reliable this number is but could swing $100M either way
• Assumed all of the costs of goods for building cars/energy products in the first two months are covered either as an accounts payable item or through the asset back line. It's possible some suppliers aren't as generous and are in a position to enforce tougher terms and some payments were already due.
• Assumed most R&D and SG&A are costs that are directly payable like salary, leases and direct costs to run the organisation. But compensated for stock based salary.
• Assuming the capex ramp up for China has not yet started in a big way
• Assuming no increase in inventory for parts to improve service as earlier promised. Could swing a few tens of millions to the downside.

Are there big ticket items that I missed?

I'd suggest reading a recent cash flow statement say from q3 to q4 to better understand the mechanics of cash flow. You have to have a view on income and go from there, adding / substracting non cash items

 

[ReflexFunds]

Which is why I assume steady state production meaning accounts payable are paid, but no new cost of goods are accounted for production following the accounts payable.



Incorrect. There is no line in my table for 'then inexplicably choosing to pay all suppliers immediately.'. Note that there is a line item for the full gross revenue value of 15k M3 and 7k S/X delivered so far with no corresponding cost of goods. There is also no cost of goods for the M3's and S/X's produced this quarter on top of those that have been delivered. That amount corresponds to the growth in accounts payable.

Please rebut or acknowledge that this was a fundamental misunderstanding of my method by you.

We can also do the exercise differently. Forget about accounts payable and accounts receivable and assume steady production. That would mean we would need to include in our table cost of goods for two months of production worth ($3.4B) as a minus on the cash position instead. To get to $2B you project, that would mean they somehow are able to stuff an additional $1,800M in the ABL and accounts payable? I am sorry, but that belies all belief.

I see what you're doing now, the cash flow line item labels are confusing but your method should largely work overall.
Looking at the numbers considering your methodology, I think the main things wrong or potentially wrong are:

• Your stock comp is subtracted rather than added back to the cash flow.
• SG&A and R&D look too high, there is also some depreciation within SG&A which should be add back. (likely $50-100m per Q).
  
• Your vehicle trade in adjustment should be positive cash flow rather than negative. Trade-ins are tied to sales volume, so if sales volume is down, clearance of used car inventory held at Dec-18 will generate much more than the cash lost from new trade-ins.
• Payables days are actually more like 70-80 on average at year end. After you remove depreciation, warranty reserve and staff costs from COGs, the actual supplier COGs is lower. This means likely $500m or so of the year end payables remain unpaid. This will actually be largely offset by COGs staff costs paid over the 2 month period, so adjusting for both of these won't change your overall numbers so much. Also on the payables side, payables days were much higher historically for Tesla than at year end, and they could likely stretch payments again in low cash flow periods.
• Your method doesn't factor in increased drawings of revolving credit and other inventory backed debt against the massive increase in new car inventory. At the minimum, Tesla will have drawn the remaining $230m on its credit line, but they likely increased this facility or temporarily added more vehicle and other loans; it is easy to get cheap debt secured by such high value car assets.
• Some of the Services COGs also have delayed payment terms, so in your method you should not have subtracted 100% of these costs from cash balance.
• Hard to know what Jan and Feb deliveries were, but I would guess your numbers are a bit low. Very wide variation between Alphahat, Insidevs, Edmunds and Bloomberg recently though.

 

[schonelucht]

I see what you're doing now, the cash flow line item labels are confusing but your method should largely work overall.
Looking at the numbers considering your methodology, I think the main things wrong or potentially wrong are:

• Your stock comp is subtracted rather than added back to the cash flow.
• ...
• Hard to know what Jan and Feb deliveries were, but I would guess your numbers are a bit low. Very wide variation between Alphahat, Insidevs, Edmunds and Bloomberg recently though.

Duh on the stock comp on my part. I do have a hard time seeing how SG&A would have $50-$100M in depreciation. What kind of items would that be? Other than that, why are they too high? I just took 2/3 of last quarter. I also think the clock is ticking on provisions for special charges on Tesla, encumbering additional capital (Elon already said they'd have one time charges for Q1 related to restructuring), especially when they plan to lay off people in Europe. Wrt the vehicle trade ins : I believe Tesla has a hard time moving inventory : the natural buyers for trade-in Teslas now have an excellent alternative in the Model 3 Medium range. Tesla moved to auction off their inventory and even then they have failed auctions. Also, they aren't just correlated with sales volume but also with resale value guarantees and expiring leases. Both most likely went up. But it is a fair criticism that they would have at least moved some of the tradins. For service I think there is very little COGS on payment terms. Just didn't seem worthwhile to make an estimate there.

All in all, I think with your corrections an upside of around $500M in total seems fair? Or would you agree otherwise?

 

[Reciprocity]

You can add 250M for SR deposits and 200M for people getting cheap FSD upgrades. That's only 100,000 of each.

 

[schonelucht]

Is there any indication that we have that kind of reservation numbers on the short range 3, let alone on the FSD upgrade?

 

[Doggydogworld]

Duh on the stock comp on my part. I do have a hard time seeing how SG&A would have $50-$100M in depreciation. What kind of items would that be?.....

Store and fixture depreciation is SG&A. Building, furniture, computers, etc. for sales, marketing, HR, legal, planning, etc. personnel. A lot of that is leased, so I don't see 100m/quarter, but 50m sounds possible.

12k+ Model 3s in the US plus maybe 1k in Canada and 5k in Europe. Only a handful in China during February. Could be as much as 20k total. On the flip side, DSO is usually around 2 weeks so you're being generous crediting 100% of AR as received without any offset for 2/28 AR.

Contrary to Reflex, I do not believe DPO for parts and supplies is 70-80 days. A chunk of AP is capex, though, which they REALLY stretch out in terms of payment. I could see 400-500m of capex payables on 12/31, of which 100-200m they may not have paid yet.

 

[SBenson]

Anyone put together a model for Q1/Q2 with the latest pricing in mind?

 

[generalenthu]

I am struggling with even delivery estimates for q1. A lot of deliveries will be in March. I am thinking 75 -80k total deliveries as my base case. We should get a poll thread out.

 

[Doggydogworld]

I am struggling with even delivery estimates for q1. A lot of deliveries will be in March. I am thinking 75 -80k total deliveries as my base case. We should get a poll thread out.

Trip Chowdhry says 30k Model 3s and 18k S/X for 48k total. Apparently Adam Jonas of Morgan Stanley also says 48k total deliveries.

Unless Tesla shut down the factory without anyone noticing, these numbers would bankrupt them. Even with the expanded ABL they don't have the cash to carry that much inventory.