What is the main challenge in scaling the production of the Model 3?

What is the main challenge in scaling the production of the Model 3?

I just read an article, via Fortune Magazine, that stated Tesla's constrain to produce the Model 3. It stated that "According to estimates from Cairn Energy Research Advisors, Tesla could ship a little over 400,000 of its Model 3 cars by the end of 2020. But before 2020, production of Model 3 could likely be constrained. For example, Tesla could ship 12,200 Model 3 cars in its first production year in 2017, and another 64,660 Model 3 cars in 2018". I was wondering if someone could answer the main challenge in scaling production? I've heard many opinions on the cause; weather it's not enough money, cost of batteries too expensive, not enough resources to produce that many battery packs, still working the details of certain features and quality over quantity. While I am sure all of those are valid reasons I have a hard time believing that those reasons are holding production as far back as 2020. If all 400,000 reservation holders bought $35,000 Model 3 cars, Tesla would have booked $14 billion in orders. That is in my opinion a pretty record setting number for a launch of a product. Money seems to be one of the issues which most people are arguing as a setback to production. Call me ignorant but I would think with that sum in orders investors would be throwing more money to alleviate most of those issues mentioned above. Another issue is the quality over quantity gospel but I believe with the recent difficulties they had with manufacturing the Model X, that they would use the lessons learned from that outcome and be able to produce a well oiled Model 3, without having setbacks all the way to 2020. I sound a bit demanding, I know, I really appreciate what Tesla has accomplished. I just don't buy the reasons to set production that far back. Perhaps I am simply just wrong but I'm just simply looking for a more sound answer.

DonS | 2016年4月28日

You mention several possibilities, and I don't doubt those are all part of the challenge, along with many more. Many of these factors are interrelated, so speeding up one part often has schedule and/or cost impacts on other items. I don't see any way to narrow it down to just one, or even a few, main challenges today.

jordanrichard | 2016年4月28日

The hold up will be battery packs from Reno and outside suppliers.

The car is being designed to be much simpler/easier/faster to build. Once the robots have been "taught" how/where to weld/rivet, it should be just a matter of speeding that part of the process, up.

It then comes to how many packs they can get from the Gigafactory and parts from outside suppliers. Unless Tesla does IMHO the smarter thing and that is build as many of the parts in house. The more they build in house, the faster they will get the parts to the production line and they will have complete control over the quality.

JeffreyR | 2016年4月28日

Cars are assembled from parts. Those parts are either designed and built by Tesla or they are purchased. So supply chain management is a major constraint. Cash flow management is super important too. That's part of the reason they don't do model years. Tesla needs to manage inventory directly. No dealerships to offload cars to.

Complex systems need to be coordinated in the factory too. Robots and people need to be trained.

Red Sage ca us | 2016年4月28日

The main challenge is production of battery cells and battery packs at the Gigafactory. Everything else about the car is pretty straightforward. ICE vehicles are produced in hundreds of thousands of units every year. But they can just order the plastic fuel tank from a supplier that likely makes millions of them a year.

It's not rocket surgery.

topher | 2016年4月28日

Can I just rant at the state of journalism?

"But before 2020, production of Model 3 could likely be constrained."
Well duh. They have 400,000 orders that they don't even have an assembly line for. OF COURSE they are 'constrained'.

And from the blindingly obvious to the totally prescient:
"Tesla could ship 12,200 Model 3 cars in its first production year in 2017, and another 64,660 Model 3 cars in 2018"
They are predicting to 4 DECIMAL PLACES how many cars they will produce two years from now!

Thank you kindly, sorry about the rant.

Drdpharris | 2016年4月28日

Crystal balls can do wonderous things :-o

Badbot | 2016年4月28日

4 DECIMAL PLACES ? Looks like whole numbers to me. unless you mean.

bb0tin | 2016年4月28日

Yup. Anyone can spread whatever rubbish they like now. Reminds me of the analyst who just said that the Model 3 would cost $US50,000 to build and Tesla would be making a loss on each one. Tesla immediately shot him down on his battery cost estimates. His projected cost was higher than Tesla's current cost. Of course that won't stop him from doing the same again.

Red Sage ca us | 2016年4月28日

topher: Heh. I stopped reading once I got to, "...Tesla could ship a little over 400,000 of its Model 3 cars by the end of 2020." I trust my own prognostications much more, and I know the ether from whence they are drawn.

PaceyWhitter | 2016年4月29日

Decimal place is another way to say significant figures. Topher is complaining about the gal to think that they can predict production with such a high level of accuracy.

Chunky Jr. | 2016年4月29日

Considering that Model 3 is supposed to start shipping at the end of 2017, doing 12,200 in that time frame would imply they can do at least 10x that many in 2018.

I expect Model 3 production to be at least 10x of what they are doing with Model X given the complexity of building that car compared to the 3.

Bubba2000 | 2016年4月29日

Tesla and Elon must be tweaking the design of Model 3 down to the millimeter or less. Optimizing aerodynamics, dash design and electronics, steering system, seats, etc. Structural design, weight and safety optimization. After they got to validate design, government tests, etc. Then order dies, jigs, modify upgrade e existing factory. Get suppliers lined up. Plus. Elon must be thinking of another factory. The critical path is the GF.

Looks to me that Tesla needs to raise $3B. | 2016年4月29日

Here are a few of the scaling Issues for 400,000 vehicles

Cash - at break even, they need 14 billion, but spread over time, less, although considerable amount must be spent before a single car is made.

Suppliers – Tesla currently has over 350 suppliers. This means in addition to Tesla scaling up, many of these suppliers have to scale as well. Just one supplier with problems could stop production. To avoid problems, Tesla has to vet each supplier and try and get multiple sources for each item.

Employees – Currently Tesla has about 13,000 employees for 80,000/yr cars. While not a linear scale, they will clearly need at many more employees – perhaps another 20,000 or so. Many of these new employees will need to be recruited, interviewed, hired and trained.

Service Centers – Needs to expand by 3-4 times the current size. This means selecting sites, buildings and having them retrofitted for Tesla’s operations. All this costs significant cash, time and effort.

Sales Centers – With a backlog, perhaps they don’t need to expand this area for a while.

Battery Factory – Need the GF to be producing the battery packs in sufficient volume. Another major expense that needs to be made in advance of sales. Reliable supplies of raw source materials must be acquired such as Lithium, Nickle, Steel, Aluminum, Cobalt and others.

Vehicle Factory – Needs a lot more people, robots, stamping machines, molding machines, casting, paint lines, QA inspection, test drivers, test tracks, storage, shipping, etc.

Supercharger Expansion - To support these large numbers of new cars, the Supercharger network must be greatly expanded, perhaps 4-6 times it's current size.

cweber | 2016年4月29日

I think Elon or some other smart dude once said something like - designing and building a machine is easy...designing and building the machines that make the machines is the hard part.

CraigW | 2016年4月29日,
Regarding the superchargers. IMO, Tesla will definitely not need 4-6 times the number of superchargers. Except in larger cities, the superchargers are very underused today (the 4-6 times usage is currently there). The Model 3 will be bought - largely - by people working for others for a living and only able to indulge in long-range travel (3+ superchargers/day) on long weekends or vacations. Therefore, I feel a noticeably lower percentage of Model 3 owners will be on the 'long distance' road than are current Model S owners, at any one time.

Superchargers are very handy, but they are still an advertising way eliminate a negative reason not to buy an EV. Perhaps the best advertising technique ever invented, but still way, way underused in actual practice. And this from someone who has used many of the different superchargers across the country.

CraigW | 2016年4月29日

I don't think anyone should minimize the problems Tesla is facing to come up to speed and produce 300-500,000 Model 3s per year. However, they also have developed the most independently automated auto manufacturing system in the world and - once the robots are properly programmed - should be able to ramp up much faster than a system almost totally reliant on people to handle most of the complex tasks.

Their problems will mostly be with distribution and supply, rather than actual manufacturing - IMO. The fact that they don't need a buffer of cars on the dealer lots and have already developed a close relationship with many businesses in the Fremont, CA area both speeds up their reaction time and creates a danger to scheduling. This is where on-site stamping and forming is crucial and multiple suppliers is required.

They are more likely to suffer parts supply problems than to be unable to ramp up production quickly. | 2016年4月29日

I should point out that I expect Tesla to handle all the issues needed to scale, but it is far from a simple task. They have some very smart people. I just don't see them producing 400,000 cars in 2018. The risks and costs are too great to expand that fast.

As for Superchargers - at least in California (where about 50% of all USA Tesla's reside) there can be busy times when all the stalls are in use at a location. Not a big problem, but these will have to be expanded for the volume of new cars that come on line. So far they have been doing a good job at expanding sites as needed. We're starting to get some locations with 12-20 Superchargers stalls. I agree that in some of the routes outside California, such as the mid-west, the usage level may remain low for quite some time.

Drdpharris | 2016年4月29日

@TeslaTap -- A very gòod outline of the challenges. This is where the mundane bites ... not as much fun as the visioning and design ... but I have faith that they foresaw this coming, but perhaps not this quick.

Red Sage ca us | 2016年4月30日

May I suggest a moratorium on the phrase, "I don't think...", please? It leaves one open for some pretty severe sarcastic comments. Thank you, All. ;-)

TeslaTap: Yes, that is a very good list, and explanation of their importance. Some might want to prioritize such a list. But really, at this point, they are all of equal importance, because the efforts on one front are useless without the others being up to the task.

CraigW: The Supercharger network might be, eventually, four-to-six times as populace as today... But only if my predictions of higher capacity battery packs allowing two-to-four times the range do not begin to appear within the next ten-to-fifteen years... Or, if faster charging technology does not appear within the next five-to-ten years. My bet is on higher capacity and faster charging. I do not believe Superchargers will be necessary at every corner, every offramp, or every 30 miles on the highway. I think it more likely that they may be available every 80-to-120 miles on popular routes, and perhaps every 160-to-200 miles elsewhere for travel between population centers. Ultimately, most people will charge at home, at least until there are a lot more used cars on the market selling to apartment dwellers in densely populated areas. I expect that for the next five years or so the majority of Tesla buyers will be those who are also homeowners. And hopefully, parking lot charging will be more plentiful at schools, libraries, condominiums, and apartment buildings within the next decade.

I do make a distinction between Capacity and Production.

I believe that Fremont must have a Capacity of no less than 300,000 units per year before Production of Model ☰ begins. That Capacity will have to increase to 400,000 before the end of Q1 2018. And it will need to increase to 500,000 by Q1 2019. So, enough to handle between 80,000 and 120,000 of the Generation II vehicles, while allowing no less than twice that amount for Model ☰.

I also believe that Production at Fremont will necessarily lag behind Capacity in order to ensure quality of the vehicles coming off the line. I hope that rate of Production is as much as 60%-to-80% of Capacity, and grows rather quickly to be within 90% at each stage of growth, before moving on to the next. 500,000 units per year is basically 10,000 per week. So starting off at a paltry 50, 100, or 200 per week will not do the trick. And with around 400,000 Reservations in place already, even matching Model S and Model X combined at 2,000 units per week will not suffice. So Production of Model ☰ must begin at or near 4,000 units per week, and grow very quickly from there. Otherwise, no one will see 'The WAIT' listed as less than two years for five years. And the earlier Tesla Motors can begin that Production rate, the better.

It will be extremely difficult to do this. Even with all the planning and logistics and contingencies, some unforeseen issues may still slow things down. So Tesla Motors must be ready to tackle those problems in short order when they arise and be flexible enough to take them in stride and move on.

topher | 2016年5月1日

In an interview Elon Musk said (paraphrased): Today's problem was 3' USB cables. Plan A supply had been shipped in a large container with stuff that caused it to be hung up in customs. Plan B supply had shipping problems. Plan C supply was sending out interns to buy every USB cable in a 20 mile radius. It is stupid things like this that cause delays. Tesla has smart folks, and Plan C's for all components but compost happens.

Thank you kindly.

JeffreyR | 2016年5月1日


I nominate @TeslaTap's answer as a candidate for copy-paste into the @OP.

Ross1 | 2016年5月2日

Haven't had time to read all the posts, but my wife has the answer.
She always does.
And is always right.

She says that while Elon is going around the factory 'picking up pop rivets' as we call it, he is the problem. He needs to separate himself from the mundane of every day and Manage.

She says Elon might be Tesla's biggest asset but he is also their biggest liability.

Ross1 | 2016年5月2日

Then again, rivets are the big problem

yongliangzhu68 | 2016年5月2日

Ross: "...She says Elon might be Tesla's biggest asset but he is also their biggest liability....

You wife nailed it, but the positives of the Tesla business model far outweighs the negatives. You will probably never see Musk using much delegated management. He is going to be in there getting his 'hands dirty' making changes and corrections until someone convinces him to allow it to ship. Of course in the end we get more innovative and interesting products because of this. | 2016年5月2日

I'd rather have Elon looking over the operation than any other car CEO. It seems like other manufacturer's CEOs only care about how they can force the engineers to make vehicles cheaper (i.e. more profitable), even if it generates more pollution, reduces safety or reduces long term reliability. I don't agree with a strategy to fix problems via advertising!

Maybe I've gotten too cynical of the car industry. Before Tesla there really wasn't much choice on how to buy a car or how cars were made. Nothing has changed other than Tesla brings a breath of fresh air to an industry too set in their ways.

Hi_Tech | 2016年5月2日

*curse curse curse*
I just spent half an hour doing an awesome write up and I get the spam filter keeping me from posting! Luckily I saved the text.
Anyone know what triggers the spam filters?

JeffreyR | 2016年5月2日

Mollom often blocks HTML, especially links and images, some strange phrases and cursing. For a list of the trigger phrases I know about see my Tips & Tricks in the General section. As you will see some of them are pretty bizarre. I think it's some kind of urban dictionary or double entendre that is involved usually. Please add any new phrases as a comment to that @OP (just replace some letters w/ underscores).
If the links are to other posts here you can keep them relative instead of fully qualified ("tips-tricks-using-and-searching-forum" vs. ""). Images are hit or miss.

Hi_Tech | 2016年5月2日

I didn't have any links, images or curses. Nice plain analysis/planning. I'll read the Tips and Tricks. Thank you for pointing it out.

Hi_Tech | 2016年5月2日

Many excellent observations above so I will not repeat those... I will however go into my expectations of the battery and vehicle factory growth plans to see if they will be the "bottleneck".

(more to come... doing it one part at a time)

Hi_Tech | 2016年5月2日

Level Setting:
a. Model 3 may begin deliveries in Q4 of 2017 with local CA owners and employees (production support period). During this time, let's say a small number of vehicles are produced and delivered to allow for quicker responses to potential issues, with minimal expenses associated to them.
b. Production ramping up for 2018 to approximately 100k (2k/week average)
c. Production ramping up for 2019 to approximately 250k (5k/week average)
d. Production ramping up for 2020 to approximately 500k (10k/week average)
e. Tesla gets enough batteries from partners to meet the needs for approx. 100k Model S and X battery packs without utilizing the GigaFactory.
f. Other model reveals and productions not specifically called out in below plan.
g. Tesla Energy demand not factored into below.

Hi_Tech | 2016年5月2日

1. GigaFactory (battery):
1a. Estimate of $5billion to complete, which includes about $1b incentives from state, $2b from partners and $2b from Tesla.
1b. Final completion will be approximately 12million sq.ft. of operational space. Will be completed with 2million sq.ft. of operational space in summer of 2016. These 2million sq.ft. spaces will be built like modules, to get to the final figure... 5 more to go.
1c. Since the fully ramped GF will be able to produce 35GWh of batteries = 500k battery packs, then each of the above "modules" should be able to create about 83k battery packs (500k/6).
1d. Tesla can build out 1 more "module" by middle of 2017, that would allow the GF to be able to produce enough battery packs for 166k Model 3 vehicles using GF battery packs only.
1e. In 2018, they can build out 2 more "modules" allowing for enough battery packs for a total of 333k.
1f. In 2019, finish completion of the final 2 "modules" (total of 500k packs capable), and complete site work of new GF2 location so that they can begin creation of "modules" at that location next.
1g. From now onwards, my back of the napkin figurers puts each module at about $500million cost (shared by Tesla and partners)

GF Conclusion - I don't see a bottleneck in terms of ramping the GF for battery packs.

2. Vehicle Factory:
2a. Here I'm slightly less aware of the costs, therefore please help correct any mistakes on my part with specifics if you have any.
2b. In 2016, Tesla is capable of producing 100k Model S and X, which is 1/5th of the overall capacity. Let's say these use two different production lines, meaning each line can produce 50k vehicles/year.
2c. Assuming it takes approximately 3 months and $250million to build out each line (baseless guesses; please correct if actuals if you are aware)
2d. In middle of 2017, Tesla sets up 3rd line for Model 3, but does limited production run for local reservations and employees of approximately 13k (1k/week for 1 quarter).
2e. Q4 2017 test of initial rollouts leads to some fixes in production line already in place.
2f. By end of Q1 2018, 2nd production line for Model 3 is setup (100k/yr or 2k/week capable).
2g. By end of Q2 2018, 3rd production line for Model 3 is setup (150k/yr or 3k/week capable). Implies 2018 capacity of 117k.
2h. Build out 3 more production lines by mid-2019, utilizing the revenues brought in from ongoing sales (total Model 3 capability of 300k/yr or 6k/week).

Vehicle Factory Conclusion - I don't see a bottleneck in terms of ramping the vehicle factory for Model 3 cars.

JeffreyR | 2016年5月2日

Great summary. Did you ever figure out what phrase Mollom had issues w/? If so, just let me know and I'll add it to the Tips & Tricks.

Josh Ensign, VP of Production at Tesla gave a great preso in the Engineering Panel at TMC Connect last summer. I am having trouble finding it, but if you poke around TMC you may be able to find it.

dd.micsol | 2016年5月2日

Wow Hi_Tech. So 400k cars would take (model 3 only) about 2.5yrs (2018/2019 and half of 2020)to produce. Hmmm. That's quite a while.
I was thinking Elon might shoot for 1.5yrs to get 400k out (rethink manufacturing). He wants to move on with a truck and other high volume lower cost EV.
Great summary though. I see what you're getting at.
Also if the gf is operational now - what's stopping them from producing batteries now and holding them so they can whip out battery packs (bp) faster (over lap) with other vendors? It's possible that they could have 150k batteries in stock by the time model 3 is in production lines. Just a thought. I don't know how many bp they can exceed over the cars they are producing right now. Maybe none. I just don't know but it is an interesting question if we all think the battery would be the biggest hold up. If they are stockpiling bp- we all could be surprised depending on how fast the factory can pump out cars. I'll leave it to the genius to figure that out. Trust him completely.

yongliangzhu68 | 2016年5月2日

dd.micsol, If I understand correctly the Giga factory can only assemble the cells into a pack and they must still be imported from Matsushita Electric in Japan. It is still a year or 2 away from actually making Lit-ion cells.

yongliangzhu68 | 2016年5月2日

Just too add they are for more expensive this way and the costs will fall substantially one they can be produced on site. Also in manufacturing you don't store up inventory. This is tax and amortization intense.

Red Sage ca us | 2016年5月2日

Nothing at all wrong with being detail oriented. There is plenty wrong with micro-managing. In Star Trek, there are two main philosophies of commanding: 1) Make It So; and 2) Get It Done. Both are very effective. But not with every crew in Starfleet.

Ross1 | 2016年5月2日

Don't forget Powerwall, it needs even more batteries than all the cars put together....

AlMc | 2016年5月2日

They need more space for production/storage: and they have it...

AlMc | 2016年5月2日

And...some news about those 'speculative reservations':

Hi_Tech | 2016年5月3日

Just a bit of clarity on my information above...
The intent was to show how they could get to the original plan of 500k/yr in 2020. By looking at this, I can see there is plenty of "buffer" built into the schedules to pull in the timelines if they want. The limiting factor would be the cash at that point... or better put, the cash flow. One thing that I hope Tesla keeps in mind is the need to become cash flow neutral/positive in 2016, while still aggressively pushing for increased production.

Hi_Tech | 2016年5月3日

@Ross - Agreed. Powerwall is point "g" in the Level Settings. I called it "Tesla Energy".
The reason I left that out is because that piece of the business should look at the finances of it separately from the auto side. | 2016年5月3日

@wj "...If I understand correctly the Giga factory can only assemble the cells into a pack and they must still be imported from Matsushita Electric in Japan."

No.The GF will take raw materials and produce the cells, and then assemble them into packs.Currently GF is only doing assembly for the Powerwall (and cells come from Japan). Fremont currently does all pack assemblies. It's unclear when the switchover will occur to using the GF for cell production and pack assembly, but perhaps starting at the end of this year.

Hi_Tech | 2016年5月3日

Tesla is assuming the GF will have this one piece of the facility fully operational this summer. Assuming that they switch over around then.

yongliangzhu68 | 2016年5月3日, I wan't saying it wouldn't do the full cell production, but that it wasn't yet doing full manufacturing of cells. I was a clearly wrong in thinking the auto pack was assembled at the GF when int is just the Powerball packs. Good to know that the GF will start full production a little early than I thought. Although if I understand (again I may be wrong) it will be very limited production at first (1 module) and ramp up but will still be about 3 years or more to reach full cell production equal to 500,000 packs annually.

Thx for clarifying the process.

Ross1 | 2016年5月4日

500,000 packs for cars. Let's say 500,000 packs the same size for Powerwall.
That is 1,000,000 packs.
If each pack has (seven kits how many were going to St Ives?) 10,000 batteries, that is 10,000,000,000.
10 billion. Divide by 365 days, divide by 24 hours, by 60 minutes, by 60 seconds.
Hang on a minute

Ross1 | 2016年5月4日

Not enough fingers.

Hi_Tech | 2016年5月4日

Let's throw in a monkey wrench.... Each PowerWall is only 6.5kwh, while the average vehicle battery pack will be about 10 times as large. So you could still have about 450k packs for cars + 500k PowerWall packs within a single GF each year.

Sorry for bringing back numbers to your light humored post. Just the geek in me. :-)

Hi_Tech | 2016年5月4日

@wj - "it will be very limited production at first (1 module) and ramp up but will still be about 3 years or more to reach full cell production equal to 500,000 packs annually."

Yes, I that is what I've come to understand as well. This first module is about 1/6th of the finished building, so assuming they would be able to make about 83k packs for each fully ramped module.

Ross1 | 2016年5月4日

I was trying to figure out how many 18650 batteries (or the next version) would be coming off the line every second.
This is truly the advantage of Tesla's multi-battery system over a mfacturer who has only 1 or 10 batteries per car. The mass production ability is enviable and unique when producing by the billions etc bla bla

Hi_Tech | 2016年5月4日

Looks like my figures are going to be completely changed with their new plan.

They may be looking to have more than 35gwh capacity at the GF1 location, if required for Tesla Energy and Autos at once. Expecting a far greater growth rate for Energy than cars.