2017 Investor Roundtable:General Discussion

[mongo]

The target is 100k deliveries of S+X, not including 3.

Good point, from Q3 letter:

Based on the recent acceleration in order growth, we now expect that Model S and Model X are on pace for about 100,000 deliveries in
2017, an increase of 30% compared to 2016. Notwithstanding these increased deliveries, we plan to produce about 10% fewer Model S
and Model X in Q4 compared to Q3 because of the reallocation of some of the manufacturing workforce towards Model 3 production. As a
result, inventory level of finished Model S and X vehicles should continue to decline.

So hitting about 100k is feasible. 27,998 are needed for 100k exactly, which is 1,057 higher than the Q3 S/X number.
So, 3 sales, reduced inventory, increased reservations, and South Australia battery revenue will be icing on the pie?

 

[Gerardf]

Good point, from Q3 letter:
So hitting about 100k is feasible. 27,998 are needed for 100k exactly, which is 1,057 higher than the Q3 S/X number.
So, 3 sales, reduced inventory, increased reservations, and South Australia battery revenue will be icing on the pie?

Norways Q4 number took care of that. It is 1375 higher than (the already great) Q3 total.
Q3 : 2359
Q4 : 3734 (and counting)

 

[brian98]

Tesla has sold in Dec more than from Dec to May 17 in 1 month
2442 and counting vs 2205

 

[Gerardf]

I added a lot of shares this month, but the share price should wait one more trading day before the big jump.

Taxes in NL are based in the mean value of investments between the dates 01/01 and 31/12. So I am fine with a big jump next week.
(Of course now I spoiled it for myself and we will see crazy jump up today, you can all thank me later).

 

[HG Wells]

Of course now I spoiled it for myself and we will see crazy jump up today, you can all thank me later

I thank you.

 

[Reciprocity]

They did say they were planning to cut S/X production by 10%, but with the guidance of 100k vehicles, that may mean a reduction in inventory.

For the quarter, 13 weeks X 2,000/W X 90% = 24,300. Or roughly 2700 cars under the required 27,000 total S/X to hit 100,000. I think this is easily offset by cars in transit, less inventory which from reports seems to be all but flushed out entirely. Cars in transit tend to be around 4-5k, so they could cut that number in half by just being way more strict on delivering vehicles farthest to closest by the end of the quarter. Certainly they will always be turning over their loaner fleet which is probably around 3000 new and CPO cars (300 locations X 10 per location). Not all locations have loaners as some are just storefronts, but that would probably be the number that they turnover every quarter.

Does anyone recall the cars in transit from last quarter? Again, we are taking about making up for 2700 cars total.

I'm general I think they are going to surprise on S/X delivers and be around 2500 model 3s. I've been watching model X vins delivery thread and some in the 79000 range. This was interesting to me because I was one of the last model x delivered last year with vin 322xx. I know it's not an exact science but that would be 47,000 model X total. With just 6k more model S deliveries (53,000) then model X, that would be 100,000. If they can drain the inventory and keep cars in transit low, then they could shock with 28,000 S/X this quarter and 30k+ total with model 3. That's my wag.

 

[Hogfighter]

If you can read the entire boring thing, the original poster comes around at the end and basically gets it. What seems to confuse bears and accountants apparently, is that cap ex is an investment that pays dividends. It's only cash burn if you have no demand for the products you're are ramping up cap ex to make. The OP babbles on about how model 3 will generate massive cash but cap ex will continue to grow because of China factory, model y and semi/roadster lines. What is missing from this thought is that each subsequent factory or model will be accelerate which means cash from those new lines will come in months quicker because they will largely be duplicates of existing successful production lines. Each line will have only incremental improvements or tweaks for model Y. Roadster and semi are not going to be huge cap ex because they will be much lower volume so they won't be hyper robotic like model 3/Y and future platforms.

The other missing piece is that the highly automated nature of the model 3 leads to lower op ex, which improves operating margins. It also leads to short production times which allows Tesla to pay for parts after the cars are delivered and payments received which will have a dramatic effect on cash flow.

A successful model 3 ramp is a game changer because the ramp to 2 million cars per year or more that follows will be based on duplication and tweaks not inventing a new way to produce vehicles. Same goes for batteries and packs. All the cars should be based on the same sub models, including semi and roadster. The minute Tesla his 10k/week, they should be ordering new equipment for the next several plants. A gigafactory is just a gigantic building with the right equipment placed in the right location. Raw materials in one side and cars and powerpacks and solar products out the other end.

The other piece of the puzzle is the software that runs the machines. When they built the first GF, they had to build the robots AND write the code that runs them. Subsequent GFs will only need the machines. Code's been written.

 

[avoigt]

This Video summarize nicely the very different perspective of comparing Tesla to a yet not existing industry which is a dynamic view to the future versus 'I compare the company to my static horizon of experience from the past extrapolating this to the future and Tesla is just another automaker competing with the big guys'. It also reminds me of the challenge of linear growth versus exponential and to get the real difference.

I struggle with an explanation why it is so hard for most people to understand the fundamental difference and disruption we experience here be it the market and demand, the new way of production, combination of industries and network effect on the financial bottom line.

You can see this in Trip desperately trying to explain it and once again after seen many of this interviews I believe all other people around simply don't really understand...

Thoughts?

 

[Reciprocity]

I just thought of something funny. There are literally 100 EV models coming by 2020. We all know that is joke and most will be scrubbed as manufactures go back to the drawing board, but let's say 20 models come out with decent production levels of 20,000/year in 2020/2021. That is roughly 800,000 non model 3s vs 1M/Y Teslas or right 2.8M potential new solar+battery customers. It will be interesting to see if Tesla can be competitive with cheap Chinese panels by then, but solar roof has no competition. 200,000 solar roofs is roughly $10B in revenue. Match that with 10B in traditional solar and batteries for residential alone. The point is, even if others manufactures are successful, they feed Tesla customers and eventually, they will dump their VW EV for a Tesla.

 

[bdy0627]

The latest update from Gene Munster/Loup Ventures is out.

They predict 2500 deliveries in Q4 and remain cautious on the Model 3 ramp (including 5000 by end of Q1) while noting positive signs. Continue to be bullish in the long run. Expecting Another Model 3 Miss, but Remain Upbeat on Tesla Story | Loup Ventures

Wow, that seems really low for Q1. I realize it is deliveries and not production, but only 5,000 for the entire quarter? That's averaging less than 400 deliveries per week. Or is he saying only another 2500 in Q1? That would be even worse. Either way, that would be hugely disappointing and also a massive miss for guidance.

 

[generalenthu]

Wow, that seems really low for Q1. I realize it is deliveries and not production, but only 5,000 for the entire quarter? That's averaging less than 400 deliveries per week. Or is he saying only another 2500 in Q1? That would be even worse. Either way, that would be hugely disappointing and also a massive miss for guidance.

Estimate for Q1 is 10.7k, 45k for Q2. Click thru to the model.

 

[EinSV]

Wow, that seems really low for Q1. I realize it is deliveries and not production, but only 5,000 for the entire quarter? That's averaging less than 400 deliveries per week. Or is he saying only another 2500 in Q1? That would be even worse. Either way, that would be hugely disappointing and also a massive miss for guidance.

Apologies but that is a typo that I thought I had fixed — should say 5K/week which Munster seems skeptical about. Exact quote is:

“Separately, the last update on Model 3 production calls for ‘a production rate of 5,000 Model 3 vehicles per week by late Q1 2018,’ which we believe is ambitious.”

 

[myusername]

This Video summarize nicely the very different perspective of comparing Tesla to a yet not existing industry which is a dynamic view to the future versus 'I compare the company to my static horizon of experience from the past extrapolating this to the future and Tesla is just another automaker competing with the big guys'. It also reminds me of the challenge of linear growth versus exponential and to get the real difference.

I struggle with an explanation why it is so hard for most people to understand the fundamental difference and disruption we experience here be it the market and demand, the new way of production, combination of industries and network effect on the financial bottom line.

You can see this in Trip desperately trying to explain it and once again after seen many of this interviews I believe all other people around simply don't really understand...

Thoughts?

This is what Trip said:

"why haven't we seen any more gigafactories? no because it is being designed by physics, first principles and only *exponential* thinkers, sadly there are only two right now, jeff bezos and elon musk who can create it"

oh... so, there's something that resembles a problem to a lot of people (lack of gigafactories)... but it's not a problem, because "physics, first principles"... and I can't understand it because "only *exponential* thinkers" can, but sadly, there's only two.

and of course, when you compare Elon to Bezos, then of course you are comparing TSLA to AMZN.

i'm sorry... Trip has always thrown my "THIS IS A MASSIVE CON" meter through the roof... and quite frankly, years ago with his inability to hide his con-like intentions was a factor in my decision to be short this stock.

at any point when someone says to me, "you don't need to understand this, you just need to believe it... now give me your money"...

I know there is something wrong.

 

[immunogold]

SP flirting with $313... I won’t miss today’s opportunity to buy!

 

[viridi]

The Munster article and the street expectations are both very odd to me. For one, Gene was just on CNBC the other day speaking highly of Tesla and saying people should invest in the company. Odd timing to then come out with a piece that will be spun by the sell-side as a negative piece from a Tesla bull (of course, from a bull perspective, I guess he is trying to be positive). Munster’s piece had a real CYA feel to it. In addition, I am not sure where the 5,200 estimate on M3 for Q4 is coming from. If Tesla really is a quarter behind, wouldn’t one expect more in the range of where they were supposed to be Q3 (around 1,500)? I know that Musk stated that production might be in the thousands per week by the end of Q4 - but, this is a run rate coming out of the quarter (which, based on what has been seen of late might be spot on). These “analysts” are really not very good at their jobs, and create so much disarray in the market place (maybe that is their goal - allows nice buying/selling opportunities for their clients). It will actually be nice to get next week behind us. We can all get actual numbers and have the speculation on Q4 production/delivery end (I know, we will then be on to the next item on which the street will speculate).

 

[bdy0627]

Estimate for Q1 is 10.7k, 45k for Q2. Click thru to the model.

Ok, so still averaging less than 1000/week deliveries in Q1. Q2 they have at an average of about 3,500/week. If I'm not mistaken Munster is typically pretty close with his estimates. I don't follow him closely but that's my impression. For those that do follow his estimates, is he usually pretty accurate? I think many here are expecting an average of about 3,500/week in Q1, which is probably optimistic. I guess the most helpful aspect of Munster's guess is that it means he believes Tesla's guidance for Q1 is wildly optimistic. EM was very clear on the Q3 CC that they had high certainty about exiting Q1 at a production rate of 5,000/week. He was equally clear that they were very uncertain about Q4 production numbers. I personally am expecting 35,000 - 40,000 produced in Q1 with deliveries obviously somewhat less than that. I would expect deliveries of at least 25,000 in Q1.

Calling Value Analyst - Could you set up a poll for Q1 Model 3 production or delivery numbers please? I'm curious what the consensus is here. Munster's prediction is making me second guess mine. I didn't think mine was wildly optimistic but perhaps I've been drinking too much Koolaid.

 

[Reciprocity]

The other piece of the puzzle is the software that runs the machines. When they built the first GF, they had to build the robots AND write the code that runs them. Subsequent GFs will only need the machines. Code's been written.

I agree in part, but if something is going to constantly change or iterate, it's going to mostly be software. To me, it's software that will mostly contribute to the 5k-10k ramp, only because I cannot figure any other way they ramp that much without large capex. The ramp will be speeding up the robots which is controlled my software and machine learning as it relates to the timings required to hit the required speeds. Things are going to be moving so fast with a high density of robots at many of the stations along the production line that they will require not just good software, but software that can adapt small variances dynamically. This could include slowdowns and bursting where variances occur across multiple stations. Almost like one large organism that adjusts dynamically to the changing flow. So I would expect the software to never be completed with thousands of changes per year to improve speed, lower error rates, improve quality above and beyond the original design. I believe Elon once stated that they made thousands of changes to S/X production over a year for this very reason.

Software is also one area that Tesla can differentiate itself as a manufacturer. It had already done so in the car but now more then ever, the software for manufacturing is going to be 10x more complex and valuable as more and more automation and speed are implemented.

Tesla was clear that going from 5k to 10k was very little capex with 100% confidence. Tesla must have a lab already running a small subset of these same processes at very high speeds to have that kind of confidence. We have seen examples of machine learning from open ai where robotic arms are trained without actually writing code. I would imagine there is some overlap between open ai and Tesla.

To me, full alien dreadnaught will require better robots. The robots will be trained like a human, only a human with thousands of arms and sensors. The robots will need to be made is space X materials with very high quality motors for extremely fine control. The physical weight of the arm and tooling will become an issue and I'm surprised it's not already at 10k/w. Also the physical size of the robot arms required to do specific Jobs. This is where the design of the car is critical as it will allow large enough openings for robots to do work inside the car. The most common example is the wiring harness. The harness must be simplified and the robots able to reach every nook and cranny inside the chassis.

Quality control is another major areas where computer vision and machine learning can improve speed and quality. We did this in the print industry with high speed cameras that took pictures of every print to verify color and adjust automatically. Every robot arm and station could have cameras taking pictures of every detail and comparing them to a huge database of perfect welds or assignments of parts. Looking for cracks or stress marks like machine learning is used today to look at x-rays and diagnose things missed by doctors.

In short, I don't think Tesla will save any capex/r&d on the software any time soon. But the software will contribute the most to long term capex savings by doubling the number of cars that can be made by the same machines, facilities and people. All of which will contribute positively to cash flow. We just have to get from here to there, which means 1k/w - 5k/w - 10k/w. Alien dreadnaught is probably a decade away but the same number of robots might be able to build 20k/w though might require Space X materials for tooling and arms.

 

[Waiting4M3]

I am also cautious on hitting 5K/wk by end of Q1 and I thought his prediction of 2500 deliveries in Q4 seemed a bit on the high side.

I guess that makes me a “conservative bull,” at least on the Model 3 ramp.

While the ramp gets most of the attention right now, I agree with Munster’s assessment that what’s on the horizon is more important than the short-term:

“Model 3 breakout in 2019. While reservation holders grow anxious and investors frustration continues to mount, we continue to stress that production over the next several quarters will be largely a guessing game, and that short-term production numbers do not materially affect the long-term story. We predict a breakout year for the Model 3 in 2019 which means, until then, other elements like solid Model S and X production numbers, increasing energy deployments like the South Australiainstallation, and future vehicles (Roadster, Semi, Model Y, and pickup truck) will stoke investor optimism.”

My Q4 expectation is more conservative but my Q1 and 2018 expectations are more optimistic. I still don't think most of the analysts know what exponential means, even bulls like Jonas and Munster.

 

[callmesam]

Hi Guys. Haven't posted in a while, but wanted the perspective from this board in particular. I’m letting the wisdom of crowds decide if I allow Jalopnik access, but It’s not looking good.

Tesla Model S60 on Twitter

I’m not trying to be cute, but the coverage of Tesla by Jalopnik has been pretty terrible. I’m a strong advocate for EVs in general, and for what Tesla is trying to accomplish in transitioning the planet to sustainable transportation.

Jalopnik . . .not so much. The feedback I've been getting is that they are bottom-feeding trolls.

If you have a different take, or there is some perspective that you personally, could share, I’m open.

Taking delivery of my first Model 3 this afternoon.

 

[Waiting4M3]

Hi Guys. Haven't posted in a while, but wanted the perspective from this board in particular. I’m letting the wisdom of crowds decide if I allow Jalopnik access, but It’s not looking good.

Tesla Model S60 on Twitter

I’m not trying to be cute, but the coverage of Tesla by Jalopnik has been pretty terrible. I’m a strong advocate for EVs in general, and for what Tesla is trying to accomplish in transitioning the planet to sustainable transportation.

Jalopnik . . .not so much. The feedback I've been getting is that they are bottom-feeding trolls.

If you have a different take, or there is some perspective that you personally, could share, I’m open.

Taking delivery of my first Model 3 this afternoon.

I have the same opinion of Jalopnik as you do, but I voted yes anyway. Maybe they're just ignorant.