Near-future quarterly financial projections

[EVNow]

I was under the impression Panasonic had said they would not be expanding production at GF1. Was that a tif, or did I read something wrong?

You read it wrong.

They wouldn't be expanding beyond 35 GWh - until they get that 35 GWh to work properly first.

 

[luvb2b]

They basically shut down solar. Deployments are down 90% since the bailout.

The solar lease portfolio continues to produce ~300m/year revenue at a 30%-ish GAAP margin.

yes but some of the income you see from solar above the operating income line is included in tesla's operating income but belongs to someone else, that's the adjustment for non-controlling interests that hits at the end. the last 3 quarters have totaled 124m wiping out half of what you believe to be the income of the solar business. you also need to back out the 53m in interest on the debt they pay each quarter, so taking that out too you'll end up same place as me: scty is a net negative contributor. i added more color below.

@neroden glad to see us agreeing a bit again. i know you didn't agree with some of my recent forecasts but they actually ended up fairly close - within 3m on gross profits (<0.1% of sales) and 15m on operating income (<0.3% of sales) for the last forecast. the first forecast for the quarter i had posted before deliveries was very close on the bottom line and within 30m on operating income (<0.5% of sales).

on a separate note, one thing that feels quite sad: if my estimates are somewhat close we would already be in the s&p 500 if we had not acquired solarcity. they've seemingly put scty into a runoff mode while waiting to launch the solar roof. look at the estimated contribution to net income since it was acquired.

i thought i understood why we acquired it a while back, but now i'm just scratching my head. no wonder they're letting it runoff. hell it's the best thing to try to just make it go away.

OK, production side, that I follow pretty carefully.

They really do believe that they've got everything going at 7K/week, and I do think the earnings release indicated that they were actually targeting higher numbers for the future. ("All manufacturing equipment in Fremont has demonstrated capability of a 7,000 Model 3 vehicles per week run rate, which we continue to work to increase.")

There don't seem to be major plans for downtime on the Model 3 lines in Q3. If they've actually reached 6500/week, that's 84500 production; even with some building up of the pipeline, deliveries should be in the 80K range. If you instead assume 2 weeks downtime (likely excessive for Q3) and 7K rate when operating, you get 77K.

So I actually do believe your 78K number for model 3, and I think it is likely to be low.

I think your S/X number is also too low. I'm expecting circa 18000.


I think your GMs are too high, though. I'd stick with same margins as Q2 on 3 and S/X. Service gross margin *needs* to drop in order to provide reasonable service.

Storage is kind of hard to calculate (we have no clear insight on the gross margins due to mixing it with solar sales & leases), but it is showing massive QoQ growth rates, and your model suggests that Q3 will be the same as Q2. With battery cell restrictions being removed, I don't see why this would be the case. I would make a WAG that they'll actually be up to 300000 in revenue rather than your 257000



OK. Could you try assuming higher Model 3 production, higher Storage sales, and lower GMs?

If I assume 15% GM on Model S/X but also assume 18K cars, this raises profit by 13350.
If I assume 18.7% GM on Model 3 but also assume 80K cars, this lowers profit by 32128.
If I assume 300000 revenue on Storage and stick with your 11% GM assumption, which seems plausible at that scale, this raises profit by 4678.50.

So I guess my model looks even worse than yours. The key issue seems to be the GM on Model 3, which makes sense. To make up for the GM staying at 18.7% Tesla would have to push out about 5500 more cars than in your model, or about 84K Model 3s. To further reach non-GAAP profitability (still assuming GM of 18.7%), they'd have to push out an additional 3000, or about 87K Model 3s (actually 86414 based on this model, but that's excess precision).

Now, that's about 6648/week. Or 7000/week with 4 days downtime. Given what I've heard... I actually think this is possible in Q3. But it's also quite possible that they'll miss the target.

It's an economies of scale business, so it's all about the volume. Either way, Q3 will have substantial positive free cash flow. Depreciation & amortization is now over $500 million per quarter, capex is under $250 million per quarter (and it was emphasized that that is sustainable), so free cash flow is running $250 million ahead of non-GAAP profit, structurally (before timing effects like inventory and accounts payable/receivable).

So they are, as they said, self-funding. They also have the cash to pay off the November 2019 bond from two quarters' free cash flow, without inventory shenanigans. They have enough cash from the stock & bond issuance to easily pay off all the non-recourse debt which comes due this year, too, and still have working cash.

TLDR: They are structurally free cash flow positive by at least $200 million/quarter. To reach non-GAAP profit they need 87K Model 3s in Q3; we'll see in early September! Higher GM would of course improve matters.

P.S. For GAAP profit, however, due to the goofy stock-based-compensation accounting, they'd need 109K Model 3s in a quarter, which seems impossible.

 

[EVNow]

I think your GMs are too high, though. I'd stick with same margins as Q2 on 3 and S/X. Service gross margin *needs* to drop in order to provide reasonable service.

We should see slightly better ASP (and margin) on S&X as most of the deliveries will be refreshed models.

They said they are getting better at selling used cars - and that accounted for better service margin. There may be some more improvement there that is possible. Infact the best idea would be to totally get out of the used car business i.e. let someone like CarMax handle all the used car trade-ins and selling those at auction (unless Tesla is actually making money on trade-ins).

 

[tmoz]

It would be nice to know how long each production item was capable of 7000 per week.
From the wording, it sounds like they haven't figured out how to get all of those pieces producing 7000 per week at the same time.
Perhaps some have spurts of only 4 hrs long, while other can achieve that rate indefinitely.
I'm sure it will be announced when they get the entire line working at 7000 per week.

 

[EVNow]

Didn't they drop the standard range battery on the S/X? That raises the ASP as much as anything.

That might have been done to make people buy inventory s/x instead. In which case, the ASP won't raise all that much, because of SR drop.

In Q2 only some 5k s/x were refreshed. Rest 12k was pre-refresh. I think in Q3 it will be the other way round. 5k pre-refresh and 12k refreshed. That should bump up the ASP.

 

[Doggydogworld]

yes but some of the income you see from solar above the operating income line is included in tesla's operating income but belongs to someone else, that's the adjustment for non-controlling interests that hits at the end. the last 3 quarters have totaled 124m wiping out half of what you believe to be the income of the solar business. you also need to back out the 53m in interest on the debt they pay each quarter, so taking that out too you'll end up same place as me: scty is a net negative contributor. i added more color below.

Oh yeah, I agree 100%. I said at the time Tesla only bought SCTY to wind it down quietly behind closed doors instead of an embarrassing public implosion that would have tarnished Musk's halo and possibly impaired Tesla's ability to raise capital.

"Take the hit on the 3 billion dollar thing to keep from harming the 50 billion dollar thing" (Adam Jonas, probably quoting someone else without attribution).

 

[luvb2b]

here's an interesting little gimmick in the letter.

Excluding regulatory credit revenue, automotive gross margin improved by ~200bp (compared to a decrease of 125bp on a GAAP basis).
​

factually this is true, however it is not fully informative.

q1 included a -92 million one time hit to gross profit, which is not being removed here.

once you take that out, the auto-excluding-credits gross margin improvement is 156 bps.

maybe it's the case that q1 also included some one time margin benefits (autopilot/fsd pull forward for example), and that's enough of an offset that the substance of the statement is meaningful.

i couldn't explain why they chose to highlight this tidbit. i wish they gave me something better.

 

[LYTRIDR]

I gotta wonder if this grand experiment will work in the end. Profit margins on new phone models is very large.(sic)

I'm more optimistic now than ever before. In my view, Tesla has a structural advantage that is just beginning to bear fruit. Look on almost every Tesla board and we see constant bemoaning of other automakers lack of interest in building a truly compelling BEV that is as good as even the 2012 Model S. The new cars from the established makers are all either too low on range, don't have the autonomous features, lack good charging infrastructure, not sexy enough, not built in enough volume to make a difference or any number of other shortfalls that limit sales of those vehicles.

This is, IMHO because of one factor. Currently, there is no one who can make a compelling BEV for the mass market at a price comperable in value to an ICE car, while making a profit. The battery-electric powertrain is too expensive to compete. Telsa could sell $100,000 cars and still lose money, but transitioning that to a $50,000 or $35,000 and making it profitable was always going to be a huge lift. The battery in the LR model 3 is ~75% as large as the Model S 100D which just last year started at $96,000 vs the Model 3 LR now starting at $48,000. That is half the price for not much less car.

Even with that huge difference, Tesla managed a FCF positive quarter, had gross margins of nearly 20%, held operations costs in check, and would have been very close to the street's estimates if not for lower regulatory credit sales and the one-time charges.

Now, Tesla has some work to do yet, but the moat they have is getting bigger. If they can continue to improve on the cost of manufacturing and in particular the cost/density and mass/density of the batteries the gross margin on the vehicles will expand. I believe that the Maxwell purchase is under-appreciated and will lead them to GAAP profits by itself, but in conjunction with the cost-savings of Giga 3 and expanded margins and market access in China, Tesla stock is looking cheap at the moment.

 

[tmoz]

I'm more optimistic now than ever before. In my view, Tesla has a structural advantage that is just beginning to bear fruit. Look on almost every Tesla board and we see constant bemoaning of other automakers lack of interest in building a truly compelling BEV that is as good as even the 2012 Model S. The new cars from the established makers are all either too low on range, don't have the autonomous features, lack good charging infrastructure, not sexy enough, not built in enough volume to make a difference or any number of other shortfalls that limit sales of those vehicles.

This is, IMHO because of one factor. Currently, there is no one who can make a compelling BEV for the mass market at a price comperable in value to an ICE car, while making a profit. The battery-electric powertrain is too expensive to compete. Telsa could sell $100,000 cars and still lose money, but transitioning that to a $50,000 or $35,000 and making it profitable was always going to be a huge lift. The battery in the LR model 3 is ~75% as large as the Model S 100D which just last year started at $96,000 vs the Model 3 LR now starting at $48,000. That is half the price for not much less car.

Even with that huge difference, Tesla managed a FCF positive quarter, had gross margins of nearly 20%, held operations costs in check, and would have been very close to the street's estimates if not for lower regulatory credit sales and the one-time charges.

Now, Tesla has some work to do yet, but the moat they have is getting bigger. If they can continue to improve on the cost of manufacturing and in particular the cost/density and mass/density of the batteries the gross margin on the vehicles will expand. I believe that the Maxwell purchase is under-appreciated and will lead them to GAAP profits by itself, but in conjunction with the cost-savings of Giga 3 and expanded margins and market access in China, Tesla stock is looking cheap at the moment.

Yes, the sky is not falling. Demand is there and their manufacturing efficiency is getting better. They are getting real world front line experience and are learning and improving from it. If model Y is as much of a hit as model 3, demand will still be greater than their production ability (a good thing actually at the moment). I think demand will have positive upward pressure from the general population getting more and more comfortable with EVs, due to a multitude of factors. So, I am also optimistic.

 

[Doggydogworld]

Didn't they drop the standard range battery on the S/X? That raises the ASP as much as anything.

They did (again), but the also cut prices (again). S/X price cuts over the past seven months have been breathtaking. P100DLs were around 150k in December vs. 100k today. 100D was high 90s, now 79,990 with more range and free AP. It's hard to see S/X ASP much above 90k going forward.

here's an interesting little gimmick in the letter.

Excluding regulatory credit revenue, automotive gross margin improved by ~200bp (compared to a decrease of 125bp on a GAAP basis).
​

factually this is true, however it is not fully informative.

q1 included a -92 million one time hit to cogs, which is not being removed here.

once you take that out, the auto-excluding-credits gross margin improvement is 156 bps.

maybe it's the case that q1 also included some one time margin benefits (autopilot/fsd pull forward for example), and that's enough of an offset that the substance of the statement is meaningful.

i couldn't explain why they chose to highlight this tidbit. i wish they gave me something better.

I think you meant to say a 92m hit to gross profit, not COGS. It was the net effect of a 500.5m reduction in Auto Sales Revenue and a 408.8m reduction in Cost of Auto Sales. I agree with the 156 bps ex-credit improvement. Q1 cost of revenues also included a 64m inventory writedown, but some of that was for used cars which hits Services instead of Auto Sales. Plus they always seem to have small inventory writedowns, so I'm not sure how much of an adjustment to apply for this one.

 

[luvb2b]

sigh...

reading this in the prior 10q:

Automotive sales revenue includes revenues related to deliveries of new vehicles, and specific other features and services that meet the definition of a performance obligation include access to our Supercharger network, internet connectivity, Autopilot, full self-driving and over-the-air software updates. Deferred revenue activity related to the access to our Supercharger network, internet connectivity, Autopilot, full self-driving and over-the-air software updates on automotive sales with and without resale value guarantee amounted to $1.04 billion and $882.8 million as of March 31, 2019 and December 31, 2018, respectively. Deferred revenue is equivalent to the total transaction price allocated to the performance obligations that are unsatisfied, or partially unsatisfied, as of the balance sheet date. Revenue recognized from the deferred revenue balance as of December 31, 2018 was $37.4 million for the three months ended March 31, 2019. Of the total deferred revenue on automotive sales with and without resale value guarantees, we expect to recognize $462.3 million of revenue in the next 12 months.​

think of the 1.04b as a "honey pot". it's money already collected that can be booked into revenue, depending on the performance of various obligations. a nice portion of this would be the autopilot and fsd sales they accumulated but haven't fully booked due to not being feature complete.

we know they launched nav on autopilot in q2. the 10q will tell us how much more deferred revenue got booked vs 19q1 (37.4m). just taking a 1/4 of the 462.3m (bolded statement abvoe) they will recognize 115m in 19q2. assuming say 70m of that is related to nav on autopilot and comes at 90% margin, it would be boosting auto gross profit by 63m or ~1.2%.

so between that 1.2% boost from one-time autopilot revenue recognition, the 0.44% boost due to including a one time adverse event in the 19q1 compare (prior note below), that's about 1.64% of the claimed 2% increase in automotive margins.

moreover, without better disclosure they can keep using this honey pot in different ways to mask what is going on with manufacturing efficiency. not too different than how the big gross margin drop in 19q1 was just washed over with the big (apparently) one-time credit sale, and then never discussed.

tl;dr: tesla is making somewhat misleading statements about gross margin efficiency by selectively using one-timers and deferred revenue.

here's an interesting little gimmick in the letter.

Excluding regulatory credit revenue, automotive gross margin improved by ~200bp (compared to a decrease of 125bp on a GAAP basis).
​

factually this is true, however it is not fully informative.

q1 included a -92 million one time hit to gross profit, which is not being removed here.

once you take that out, the auto-excluding-credits gross margin improvement is 156 bps.

maybe it's the case that q1 also included some one time margin benefits (autopilot/fsd pull forward for example), and that's enough of an offset that the substance of the statement is meaningful.

i couldn't explain why they chose to highlight this tidbit. i wish they gave me something better.

thanks i edited the post to correct it.

They did (again), but the also cut prices (again). S/X price cuts over the past seven months have been breathtaking. P100DLs were around 150k in December vs. 100k today. 100D was high 90s, now 79,990 with more range and free AP. It's hard to see S/X ASP much above 90k going forward.


I think you meant to say a 92m hit to gross profit, not COGS. It was the net effect of a 500.5m reduction in Auto Sales Revenue and a 408.8m reduction in Cost of Auto Sales. I agree with the 156 bps ex-credit improvement. Q1 cost of revenues also included a 64m inventory writedown, but some of that was for used cars which hits Services instead of Auto Sales. Plus they always seem to have small inventory writedowns, so I'm not sure how much of an adjustment to apply for this one.

 

[M3BlueGeorgia]

Just listened to the Earnings Conference call (Q2-19)

Interesting, but predicable, that Elon is warning that Q1-20 will be a bad quarter. That's always a really bad quarter for the automobile industry, but most manufacturers hide it by stuffing inventory onto their dealers lots. Tesla can't do that, because of their direct sales methodology.

It'll be interesting to see how they handle that situation in 6 months time.

My recommendation is that January 2020 is a really good time to launch the SR+/AWD

 

[EVNow]

so between that 1.2% boost from one-time autopilot revenue recognition, the 0.44% boost due to including a one time adverse event in the 19q1 compare (prior note below), that's about 1.64% of the claimed 2% increase in automotive margins

How much of AP/FSD rev did they recognize in Q1?

The other way of looking at this is, they are spending a lot on FSD computer, sensors, r&d etc. That's already in cogs.

If this recognition becomes ongoing, then it will won't be a one time margin boost. Esp if they can recognize most of FSD revenue at the time of sale. I expect that to be the case after FC in Q4.

 

[M3BlueGeorgia]

How much of AP/FSD rev did they recognize in Q1?

The other way of looking at this is, they are spending a lot on FSD computer, sensors, r&d etc. That's already in cogs.

If this recognition becomes ongoing, then it will won't be a one time margin boost. Esp if they can recognize most of FSD revenue at the time of sale. I expect that to be the case after FC in Q4.

If I were running Tesla, I'd try to recognize more revenue in Q1-19 from FSD, but that would require delivering functionality which is still in development, so doesn't seem likely that they can recognize all FSD revenue by then, but maybe half?

Elon also indicated that most owners haven't signed up for FSD. Not a big surprise. So there's also a revenue opportunity to upsell existing owners once there is more capability available.

 

[Doggydogworld]

How much of AP/FSD rev did they recognize in Q1?

The other way of looking at this is, they are spending a lot on FSD computer, sensors, r&d etc. That's already in cogs.

If this recognition becomes ongoing, then it will won't be a one time margin boost. Esp if they can recognize most of FSD revenue at the time of sale. I expect that to be the case after FC in Q4.

They don't say, but the 37.4m of previously deferred revenue they recognized in Q1 is consistent with prior trends which only included Supercharging, connectivity and such. So I'd guess they recognized zero previously-deferred FSD. Post-2/28 "new FSD" sales include "old EAP" features like AutoPark and Summon for which they can recognize revenue immediately, of course. I don't know what portion of New FSD sales they recognize immediately but I'd guess 20-30%.

 

[M3BlueGeorgia]

You read it wrong.

They wouldn't be expanding beyond 35 GWh - until they get that 35 GWh to work properly first.

Yep, but they are now up to around 28 GWh, so starting to close onto capacity limitation.
Its going to be a rush to see whether they get to full capacity before they start putting an alternative battery production systems in place.

In the Earnings call they talked about needing eventually 100 fold increase in battery capacity.

 

[neroden]

yes but some of the income you see from solar above the operating income line is included in tesla's operating income but belongs to someone else, that's the adjustment for non-controlling interests that hits at the end. the last 3 quarters have totaled 124m wiping out half of what you believe to be the income of the solar business. you also need to back out the 53m in interest on the debt they pay each quarter, so taking that out too you'll end up same place as me: scty is a net negative contributor. i added more color below.

@neroden glad to see us agreeing a bit again. i know you didn't agree with some of my recent forecasts

I remember saying I couldn't predict Q2 and wasn't going to due to the extremely high uncertainty -- don't take that as disagreement! I'm impressed you were able to make decent forecasts.

but they actually ended up fairly close - within 3m on gross profits (<0.1% of sales) and 15m on operating income (<0.3% of sales) for the last forecast. the first forecast for the quarter i had posted before deliveries was very close on the bottom line and within 30m on operating income (<0.5% of sales).

on a separate note, one thing that feels quite sad: if my estimates are somewhat close we would already be in the s&p 500 if we had not acquired solarcity. they've seemingly put scty into a runoff mode while waiting to launch the solar roof. look at the estimated contribution to net income since it was acquired.

View attachment 434072
i thought i understood why we acquired it a while back, but now i'm just scratching my head. no wonder they're letting it runoff. hell it's the best thing to try to just make it go away.

Buffalo Factory carries a massive pile of extremely generous state tax breaks and subsidies and was too good to lose -- it was undervalued. (It has not paid off yet; it will.)

Plus, the existing solar leases were worth more than they were being valued at, so they were being bought at essentially a bargain price. They were promptly cashed out using asset-backed notes and similar, which is, I believe, why the income on them is now cancelled by the interest payments on the debt backing them -- it was actually a stealth cash raise.

The solar lease accounting is SO screwy that it's not going to resolve to cash until 30 years are up. I think looking at cash flow may give more of a clue as to why it wasn't such a bad deal. The solar lease accounting is such a cloud of fog -- I dug through it all in detail ONCE when the merger was happening, and I never ever want to do it again. I concluded at the time that it was structurally profitable, but that the cash outlays were immediate and the payback was late. Tesla reversed this by securitizing all the existing leases to get the cash back; with Tesla's backing they were able to get pretty good deals and so I think they ended up making a cash profit on these. But it has NOTHING to do with GAAP accounting, which is screwball in this area. I'm very glad solar leasing is ending and it should have stopped sooner, if only because the accounting is quite incomprehensible.


I think part of the rest of it is reputational; Musk couldn't afford to let the company go bankrupt, so when he saw the opportunity to buy it for less than its runoff value... he took it. He also definitely wanted the commercial/utility-scale installer teams, (which he mentioned specifically during the acquisition), and they are still being used.

In short, I still think it ended up being a good deal, although they should have ended solar leasing *faster* and should have eliminated the retail sales teams *faster*. The merger generated non-GAAP-reportable value. The liquidation of the no-good zero-money-down leasing and the no-good expensive sales channels took years longer than it should have, however.

 

[neroden]

It would be nice to know how long each production item was capable of 7000 per week.
From the wording, it sounds like they haven't figured out how to get all of those pieces producing 7000 per week at the same time.
Perhaps some have spurts of only 4 hrs long, while other can achieve that rate indefinitely.
I'm sure it will be announced when they get the entire line working at 7000 per week.

The interesting point to me is that they are clearly targeting even higher rates. That's a good sign.

 

[EVNow]

Here is my P&D estimate - as well as P&L. This is what is needed for break-even in Q3 & profit in Q4, which I think are possible.

 

[M3BlueGeorgia]

Here is my P&D estimate - as well as P&L. This is what is needed for break-even in Q3 & profit in Q4, which I think are possible.

View attachment 434295

View attachment 434296

IMHO you are underestimating sales of S and X in Q3, and therefore in Q4-19.
Perhaps add 25% for both the main automotive sales season and the benefit of the Raven refresh, so an extra 4K units each quarter.

Your production estimate for Model 3 in Q4-19 doesn't appear to take account of the China factory coming online,
So say add 1500/wk for half that quarter, even though Tesla seem to think they might get to 3K by the end of the quarter.
So say 7*1.5= 10K vehicles from China factory in Q4-19 raising total Q4 Model 3 production to 95K, and that should add directly to sales without increasing inventory.

Of course Q1-20 is the big challenge. Since China factory just came online with a lot of immediate non-seasonal demand, you can assume a full contribution of (say) 30K vehicles from it, both production and sales.
Elsewhere in the Northern Hemisphere, Jan-Feb is a really lousy time to be selling cars (and March isn't great), so Tesla should have a mild end-quarter inventory build up, and some drop-off in sales (from Q4-19). So perhaps take Q1-19 sales and add 30K China Model 3 to the mix, and call that a forecast?