2017 Investor Roundtable:General Discussion

[erthquake]

Thank you for challenging me with a reasonable argument. I highly appreciate it as this helps me tremendously.

The following are the reasons why I see Tesla moving forward simultaneously and quickly on "Gigafactories 3, 4, and possibly 5."

1. Looking at investing (and other aspects of life) through a Game Theory lens has helped me tremendously over the years. Why would Tesla announce locations for possibly three additional Gigafactories at the same time, if they did not have the intention to break ground shortly thereafter (within three months as they did with GF1) and build all three simultaneously? Bears can say maybe Elon is trying to manipulate the stock, but I don't see that in his history despite his tweets and for the following reasons.

2. The demand is there (skip this paragraph if you agree). Tesla has captured one-third of the US luxury sedan market in four years, extremely quickly in terms of automotive industry timelines. Model X is on track to achieve that market share level in less than two years! I do not see a reason why the Model 3 cannot achieve the same level, other than production capacity. Even if the Model 3 eventually ends up with one-tenth of the market share the Model S/X achieved with smaller number of superchargers, smaller number of service centers, and lower brand awareness than what Tesla enjoys today, Tesla will not be able to meet demand, even with five Gigafactories. I hope we agree that demand is not the limiting factor for my prediction to be realized.

3. You raise a good point about capex needs to achieve such a goal, but if one actually thinks about how much capital is needed, and compares that to the benefits, one quickly realizes that (given that demand is there) Tesla should be building as many Gigafactories as possible. Basically, starting a Gigafactory costs a few hundred million dollars and a few billion over the full course of the project with majority of the cost being in tooling (so towards the end of the construction project). So the cost of starting all three simultaneously would be maybe a billion dollars. If the Model 3 ramp-up goes smoothly, as it seems to be so far, and Tesla produces 1 million cars over the next ~20 months, which I think is a strong possibility, that's $50B+ of revenue. Even if the gross margin is ~15% in earlier days of the Model 3 ramp up, that's many more billions than what it would take to start the Gigafactories 3, 4, and possibly 5. Further, I expect Tesla to now be able to take on non-dilutive corporate debt as its revenue multiplies in the next 12 months. This simple back-of-the-envelope calculation looks even brighter for Tesla beyond 2018 as the Model 3 production line (and gross margin) reaches maturity. Much much brighter.

4. So we established that the demand is there and that the cap-ex needs in the early construction stages are very manageable with an even brighter outlook in later years. What is the benefit? The benefits are two-fold: 1) the acceleration of production capacity pulls forward tens of billions of cash flow, and to any trained eye that has looked at discounted cash flow models (and I've worked on hundreds), such a move increases the Enterprise Value of a company tremendously. More importantly, if Tesla can take market share from incumbents quickly, the probability of ICE car manufacturers going bankrupt before they can build their own Gigafactories (which is a must to compete with Tesla) increases. Again, this is another example where thinking strategically (i.e. game theory) is more helpful than just purely in numbers.

5. Finally note that the full build-out of Gigafactories 3, 4, and possibly 5 is not needed to realize my forecast of 4 million cars in 2020. With GF1 capacity raised to 1.5 million cars (triple of original forecast just two years ago), Elon/Tesla's history of pushing limits by multiple times and "order of magnitude" improvements version after version, recent creation of Tesla Advanced Automation, Elon's repeated comments around becoming the best manufacturer in the world, I believe even a partial build-out of the three additional gigafactories three years into the construction projects would be enough to achieve the 4 million car total capacity among the four Gigafactories.

I hope this is helpful. Please keep challenging me with thoughtful reasoning. I very much appreciate it.

What's your best guess on raw material availability for batteries? Cobalt seems to be the most likely constraint, but I don't know about how much Tesla has been able to secure for future production. This would be helpful information to make sure your forecast of 4M cars in 2020 is viable.

 

[FredTMC]

So, I was a bit optimistic when I posted about the 467 robots versus the 160 in the Fremont factory. I guess that 160 number was in 2013. The Model X line has 542 robots on its production line.

As a side-note, some people are discussing that if these robots haven't been installed yet, how is it that the M3 CR cars were made using production hardware? They probably ran them on the MS or MX line, testing out the simulated runs of the hardware that would be the same using different programs, allowing them to make rough adjustments early on. The changes then can be applied to the equipment that is being installed now.

Thx

Couple points:

Please reference which type of "Assembly line" in discussions (this is directed at everyone). Body-in-white, Subassemblies, or Final Assembly.

There is zero chance Tesla built the few M3 RCs on the MX/MS Final Assembly Line (IMO).

Likewise there is little chance they used the MS/MX BIWs Robotic weld lines. Those are fully utilized

 

[Lhan]

I really liked your post a lot. Thanks for sharing.

I'm still completely struggling with where these 476 robots are getting deployed:

Here's my guesses:

1). 200 ROBOTS. Body-in-white line is nearly 100% robotic. I assumed BIW line was ALREADY in place and used to build RC car BIWs. Apparently not given how many robots got delivered. I suppose it's possible tesla is building out a second M3 BIW line after validating the first Model 3 BIW line

I think the Newer X/S Robotic BIW line uses MORE robots that the original BIW line. Tesla determined that it was not efficient to have the robots switch "hands (tools)"


2). 200 Robots Subassembly Robotic lines. Similar to the MX robotic video posted yesterday on a Electrek that does work on an MX fender quarter panel (I think).
I can imagine there are more robotic subassemblies for M3 than we're ever done for S/X


3). 50 Robots . Final Assembly line. At this stage of M3 ramp, I can't imagine tesla will have more than a single Final Assembly Line. But I agree there could be stations that are parallel. For S/X, there are few stations that are robotic such as Roof and Tires. I assume for Model 3 there are a few more stations that are Robotic but still that vast majority of Final Assembly is human.

4). 26 Robots . Misc areas. Stamping, station transfer, Paint, etc.

Would it not be more effective to add duplicate stations to a linear workflow, and divide the more time consuming work between then, so not to slow the line down ?. instead of merging a car to and back from a parallel line ?.

 

[FredTMC]

Would it not be more effective to add duplicate stations to a linear workflow, and divide the more time consuming work between then, so not to slow the line down ?. instead of merging a car to and back from a parallel line ?.

I assume you're referring to the "final assembly line"

We have no idea how much parallelism will be in some of those stations

I don't recall there being any parallelism in any stations on the MX/MS lines

 

[S3XYTSLA]

twitter is "just ok". stock is heavily shorted and barely even up 10% on a great report.
what you want to see is a large heavily shorted stock up 20-40% on a good report. that both strikes terror in the hearts of the shorts and emboldens speculators to press against their other hideouts.

as i mentioned in my earlier post, the action today might lead you to conclude upside is as much as to 340. and that's nice but "just ok" because it's very hard to trade for extreme profits with options.

not sure if i remember correctly, did you say TSLA needs 2 quarters in a row for S&P considerations.. i read it was 4...

 

[MitchJi]

1. I assume you're referring to the "final assembly line"
   
   We have no idea how much parallelism will be in some of those stations
   
   I don't recall there being any parallelism in any stations on the MX/MS lines

   On the Q4 ER Elon's statements either specifically state or strongly imply parallelism on the alien dreadnaught line when ramping from 5k to 10k per week.
   

 

[ValueAnalyst]

I don't have any question about demand being sufficient to support multiple GFs, but a few thoughts:

• Tesla has raised about $2.7B in capital in 2016 and 2017 just to get to the start of Model 3 production. Additional capital may be required for the ramp to 350-400K Model 3s per year in 2018. The capital required to go from 500K vehicle production to 4M is likely to be massive, especially if done on the timeline you suggest.
• I believe Elon's actions clearly demonstrate he is deeply averse to both dilution and debt. Dilution means the loss of voting power which is a potential major threat to Tesla's future, and excessive debt (as viewed by the market) could make Tesla vulnerable to attacks by shorts ala Solar City. Hence the convertible debt offering this year structured to minimize risk of dilution, and Elon's continued TSLA purchases.
• Groundbreaking on GF1 was June 2014. Tesla Gigafactory | Tesla Assuming Tesla completes it in 2020 as predicted, the full build will have taken 6-plus years. Obviously the next Gigafactories need not take this long, but this extended timeline is worth balancing against the idea that the only reasonable path forward once GF locations are announced is a very rapid build.
• As others noted above, future GFs will include not just batteries, drive trains and TE products but vehicle production.
• Tesla's technologies, especially production technologies, are likely to continue to advance rapidly. Building out multiple GFs with Alien Dreadnaught 1.0 technology may not be the most efficient path forward.
• TE and Model 3 (in addition to S, X and possibly solar products) should generate cash but how fast and how much is uncertain.

Based on these factors and the uncertainties built into them, it is difficult for me to estimate the "optimal" timeline to build GFs 3, 4 and 5. I am highly skeptical that Tesla will ramp GF buildouts and reach 4M vehicle production as fast as you suggest. On the other hand, I believe the GF ramps and new product rollouts will happen much faster than the market expects, and that exceeding 1M vehicles per year in 2020 is likely.

Glad to learn that we agree on both (i) demand is sufficient to support multiple Gigafactories and (ii) exceeding 1M vehicles per year in 2020 is likely. The following are my responses to your points in order:

• I agree that large amounts of cap-ex will be needed to get from 500K to 4M, although please note that both Elon and JB pointed out that the ratio of additional cap-ex to additional capacity is not a 1:1 ratio, which makes sense. What I'm also saying, however, is that Tesla will have tens of billions coming in 2018, as well as further access to non-dilutive debt, so I fully expect Tesla to finance additional Gigafactories with internal cash flow and non-dilutive debt, as opposed to further equity raises at $300/share or convertible debt.
• Given the large margin with which the seemingly controversial SCTY acquisition was approved, and its decade-long relationship with largest investors, there's a long way before Elon has to worry about dilution of voting power. I do, however, agree that he seems to be aligned with investors with respective equity value dilution, at least at $300/share. I expect Tesla to shift to non-dilutive corporate debt to finance further expansion as total net debt to enterprise value ratio is extremely little for, technically, a large-cap company.
• Tesla broke ground in Nevada on August 1, 2014 with the aim of ramping up to 500,000 cars by 2020, which was later pulled forward to 2018, because significantly more than expected demand for the Model 3 allowed Tesla to tap the capital market. So four years from breaking ground to achieving initial max production goal. Because I expect Tesla to finance additional Gigafactories primarily from Model 3 gross profits, I expect it capital constraints on production timeline to lessen even further going forward. I think we'll have to agree to disagree on the timeline for additional Gigafactories until additional management commentary and data on Model 3 demand comes out in the coming months.
• I would argue that bringing parts assembly under one roof with battery production may actually significantly increase the pace at which Tesla can produce cars per Gigafactory. Tesla is currently shipping batteries from Nevada and Fremont via rail. Imagine having a Whole Foods in your kitchen; would it be quicker or slower for you to go from shopping to creating the delicious dinner that I'm sure you cook?
• Alien Dreadnought 1.0 is GF1 in summer 2018, so I expect the next-gen Gigafactories 3, 4, and possibly 5 to be Alien Dreadnaught 2.0, so I actually expect higher than 1.5 million cars/year max capacity for these Gigafactories, but this is not part of my 4 million cars/year run-rate by 2020 "base case" forecast. Gigafactories 1, 3, 4, and 5 with Alien Dreadnaught 1.0 speed would be enough for my forecast. I don't think the TMC community is ready for my "best case" forecast...
• Model 3 should get to +15% margin by 1Q18 assuming no Model X-like ramp-up issues. I also expect Model S/X margins to improve in 2H17 and into 2018 as Gigafactory volume production ramps up and as Model X achieves even higher sales rate. There may also be some purchasing power related cost efficiencies for Model S/X due to significantly higher volume Model 3 parts purchases. None of this is included in my "base case" scenario.

I appreciate you following up with further thoughts. I think we have likely reached the point at which we need further commentary/data to fine-tune our forecasts, but I'm glad we now fully understand the specific points at which our forecasts take slightly different routes.

 

[Reciprocity]

not sure if i remember correctly, did you say TSLA needs 2 quarters in a row for S&P considerations.. i read it was 4...

I Luvb2b said before its 4 total quarters of profitably on average, so you need this quarter and next to be profitable if you include q3 and q4 last year. But I could be wrong.

 

[FredTMC]

1. 
   On the Q4 ER Elon's statements either specifically state or strongly imply parallelism on the alien dreadnaught line when ramping from 5k to 10k per week.

Of course. But that's in the future for scaling

I'm talking about initial launch

 

[EinSV]

Glad to learn that we agree on both (i) demand is sufficient to support multiple Gigafactories and (ii) exceeding 1M vehicles per year in 2020 is likely. The following are my responses to your points in order:

• I agree that large amounts of cap-ex will be needed to get from 500K to 4M, although please note that both Elon and JB pointed out that the ratio of additional cap-ex to additional capacity is not a 1:1 ratio, which makes sense. What I'm also saying, however, is that Tesla will have tens of billions coming in 2018, as well as further access to non-dilutive debt, so I fully expect Tesla to finance additional Gigafactories with internal cash flow and non-dilutive debt, as opposed to further equity raises at $300/share or convertible debt.
• Given the large margin with which the seemingly controversial SCTY acquisition was approved, and its decade-long relationship with largest investors, there's a long way before Elon has to worry about dilution of voting power. I do, however, agree that he seems to be aligned with investors with respective equity value dilution, at least at $300/share. I expect Tesla to shift to non-dilutive corporate debt to finance further expansion as total net debt to enterprise value ratio is extremely little for, technically, a large-cap company.
• Tesla broke ground in Nevada on August 1, 2014 with the aim of ramping up to 500,000 cars by 2020, which was later pulled forward to 2018, because significantly more than expected demand for the Model 3 allowed Tesla to tap the capital market. So four years from breaking ground to achieving initial max production goal. Because I expect Tesla to finance additional Gigafactories primarily from Model 3 gross profits, I expect it capital constraints on production timeline to lessen even further going forward. I think we'll have to agree to disagree on the timeline for additional Gigafactories until additional management commentary and data on Model 3 demand comes out in the coming months.
• I would argue that bringing parts assembly under one roof with battery production may actually significantly increase the pace at which Tesla can produce cars per Gigafactory. Tesla is currently shipping batteries from Nevada and Fremont via rail. Imagine having a Whole Foods in your kitchen; would it be quicker or slower for you to go from shopping to creating the delicious dinner that I'm sure you cook?
• Alien Dreadnought 1.0 is GF1 in summer 2018, so I expect the next-gen Gigafactories 3, 4, and possibly 5 to be Alien Dreadnaught 2.0, so I actually expect higher than 1.5 million cars/year max capacity for these Gigafactories, but this is not part of my 4 million cars/year run-rate by 2020 "base case" forecast. Gigafactories 1, 3, 4, and 5 with Alien Dreadnaught 1.0 speed would be enough for my forecast. I don't think the TMC community is ready for my "best case" forecast...
• Model 3 should get to +15% margin by 1Q18 assuming no Model X-like ramp-up issues. I also expect Model S/X margins to improve in 2H17 and into 2018 as Gigafactory volume production ramps up and as Model X achieves even higher sales rate. There may also be some purchasing power related cost efficiencies for Model S/X due to significantly higher volume Model 3 parts purchases. None of this is included in my "base case" scenario.

I appreciate you following up with further thoughts. I think we have likely reached the point at which we need further commentary/data to fine-tune our forecasts, but I'm glad we now fully understand the specific points at which our forecasts take slightly different routes.

I won't go through a point-by-point response but a couple high level reactions. IMO you are underestimating Elon's desire to avoid future dilution, as well as Tesla's aversion to taking on significant amounts of debt, so we can agree to disagree about that. Also, given the most ambitious guidance provided by Elon is 1M vehicles in 2020, your 4M in 2020 "base case" estimate puts you way out in front of company guidance.

I also suspect the timing of the plan to announce the locations of GF3, 4 and possibly 5 has more to do with the need to reduce costs and avoid currency issues by moving production closer to the location of delivery, as well as the need to support production of battery hungry vehicles like Tesla Semi and Pickup. I agree the added capacity will also be needed to ramp the overall volume of vehicle production well beyond 1M per year but IMO that will happen a bit slower than you are predicting.

I do think it is interesting to think through all the possibilities, so thank you for raising.

 

[ValueAnalyst]

This is an important paragraph from Elon's recent e-mail to Fremont employees. It shows, as I had thought, that Elon is even more bullish than the "ten-fold increase in the next 5-10 years" included in his e-mail to Grohmann employees.

"The chart below contrasts the total comp received by a Tesla production team member who started on January 1, 2013 against the total comp received over the same period at GM, Ford, and Fiat Chrysler (FCA). A four year period is used because that’s the vesting length of a new hire equity grant. I believe the equity gain over the next four years will be similar."

I agree.

 

[FredTMC]

Well, April is almost over... waiting for Solar Roof ordering launch (and pricing)

Tick Tock, Elon. You said April and I'm holding you to it.

 

[luvb2b]

the way i read it the last 4 quarters total need to show gaap profit and the most recent quarter needs to show a gaap profit. if they come in with a gaap profit this quarter as i expect, a modest q2 gaap profit should get them to meet all the criteria.

this was actually an impossibility based on their operating plan if not for the solarcity acquisition (how ironic is that!). the inclusion of massive solarcity nci's could easily result in back to back gaap profits, which is why i beat that sorry plug senseless.

not sure if i remember correctly, did you say TSLA needs 2 quarters in a row for S&P considerations.. i read it was 4...

 

[vgrinshpun]

LG Chem launches their residential battery lineup (RESU) in US.

 

[geneclean55]

twitter is "just ok". stock is heavily shorted and barely even up 10% on a great report.
what you want to see is a large heavily shorted stock up 20-40% on a good report. that both strikes terror in the hearts of the shorts and emboldens speculators to press against their other hideouts.

as i mentioned in my earlier post, the action today might lead you to conclude upside is as much as to 340. and that's nice but "just ok" because it's very hard to trade for extreme profits with options.

Not sure if TWTR earnings report was great. It seems like the losses were less than expected. For example, revenue was better than expected but still in a declining trend.
Also, analyst consensus had a +1 cent EPS. TWTR ended up posting 11 cent EPS.
Analysts are negative EPS for TSLA. Is that right?

 

[erthquake]

LG Chem launches their residential battery lineup (RESU) in US.

No pricing info, but still significantly lower capacity compared to Powerwall 2, and no included inverter.

 

[EinSV]

LG Chem launches their residential battery lineup (RESU) in US.

Any info available on pricing?

 

[racer26]

not sure if i remember correctly, did you say TSLA needs 2 quarters in a row for S&P considerations.. i read it was 4...

The sole S&P inclusion requirement Tesla does not currently possess is to have the sum of the trailing 4 quarters be positive as well as the most recent quarter.

 

[vgrinshpun]

Any info available on pricing?

I looked it up a while ago, and based on Australian market I believe it was higher than Powerwal 2 (on a per KWh basis) I do not have time to dig it out, but comparison is available from the Australian internet Sites.

As far a I remember, Powerwall also has wider ambient temperature range...

 

[DurandalAI]

I would also make a wild guess and say that Tesla has validated programming of the robots to do every step of the assembly during the RC build. Obviously they have not put in all the robots needed to run the line at a high volume given the pictures from yesterday. And I would also guess that the robots may have been ran in manual mode during the RC build, and Tesla is doing copy/paste and setting up automation now.

Here is an example of how we introduce new production line and add capacity in semiconductor fabs:

We would 1st install 1 tool, run verification/calibration (usually quick). Then we would run boxes of wafers through it manually, to tune the processing recipe (think Tesla's robot programming). The recipe tuning is usually the most time consuming, you don't always know if the recipe works, and you're not done until you end up with a recipe that works.

Once the recipe is done, we set up the automation. This means programming the wafer conveyor to transport wafers to the tool automatically, and set up communication between the tool and the fab automation server, so no human needs to be present to load the wafers and hit a Start button. This part is quick, especially if you have done this for other tools in the fab, maybe on a previous production line, you just copy/paste. Now tool #1 is online and production has started.

Then we add the next tool. 1st we also do verification/calibration (quick), then copy/paste the same recipe (much quicker than tuning recipe), then run a box of wafers and verify the results, then we copy/paste the automation and tool #2 is online.

Then we add the next tool, and the next tool, and the volume capacity of the line increases gradually as we add more tools.

What you're describing is what I was leading towards with my early post, that they probably used robots, possibly ones that are on the S or X lines during the 2 week shutdown to do the RC build. Otherwise, I don't think that Elon Musk's predictions of 1,000/week only a month in would be possible, let alone the 5,000/week before the end of 2017. The closer we get to July, the closer I really think Tesla will meet or very barely miss its projections on M3 production unless "stuff happens."