Cybertruck Production Ramp Estimates

[AlabamaMike]

I am attempting to add a column to the Cybertruck Reservation tracker that estimates how long someone will have to wait to receive their order after an eventual factory starts production. The tracker is here:

Cybertruck Reservation Tracker

If you scroll up to the top and right in the google sheet you will see the start of assumptions on ramp up. As I worked on estimates, I quickly realized Tesla is in a tight spot. Even assuming rapid ramp up, Tesla is at least 4 years away from catching up. Here are some key assumptions:

• Tesla will start with Tri Motor and move to Dual then Single Motor as economies of scale improve.
• A potentially large number of reservations will cancel out of frustration or may not have the ability to follow through
• Even though the Cybertruck will require considerable new methods of production (along with their own production h*ll), tesla will ramp up quickly: 100k+ within 12 months; 250k+ during second year; 500k+ during year 3 and 4.
• We are 18 months away from the first unit being produced for sale. (6 months to ground breaking; 12 months after that to deliver 1st truck).

What do people think about above? feedback?

As far as reservations:

• reservation rate will quickly hit a steady state of 10% of the US truck market (about 240k a year). Right now the reservation rate is running around 600k a year so this is conservative.
• reservations will continue to build however 25-35% of those will eventually cancel. Does anyone know how this compares to model 3 cancellations?
• Tesla will have to go for it as far as maxing out capability (500k per year) in order to catch up with demand in the next 5+ years; after that they will have to adapt some lines to additional Cyber Vehicles (maybe vans? or cyber cars?) to scale back to a steady state of around 240k per year. If Tesla simply tries to asymptotically ramp to demand it will take 10 years to catch up and people will cancel orders in droves.

What do people think? feedback?

Here is the current state of my uneducated predictions (you can explore the formulas used on the google sheet):

	Tri Motor	Dual Motor	Single Motor	Total	Cumulative				Est Reserv 18m from now
Monthly Rate	25%	30%	35%	<- Additional cancel	288,176	307,969	42,166	638,311	889,061	18 month from now
0-3 month	3,000	0	0	3,000	9,000	0	0	9,000	<- Total Production
3-6 month	6,000	1,000	0	7,000	27,000	3,000	0	30,000
6-9 month	10,000	4,000	0	14,000	57,000	15,000	0	72,000	Total Reserv
9-12 month	15,000	8,000	0	23,000	102,000	39,000	0	141,000	875,561	30 month from now
12-24 month	15,000	15,000	0	30,000	282,000	219,000	0	501,000	1,112,811	42 month from now
24-36 month	16,000	19,000	7,000	42,000	474,000	447,000	84,000	1,005,000	1,350,061	54 month from now
36-48 month	16,000	19,000	7,000	42,000	666,000	675,000	168,000	1,509,000	1,587,311	66 month from now
mix on final year ->	38%	45%	17%		44%	45%	11%	<- cumulative mix

Looking for feedback.

For example we currently estimate over 500k reservations given an assumed cancellation rate and over 600k raw reservations. In 18 months I estimate they will explode to 889k reservations. Between 25% and 35% of those will not follow through (depending on Trim package - see above). This leaves 640k reservations to fill. Tesla will ramp up Tri Motor first to account for low economies of scale and satisfy those Tri Motor reservations within 2 years. However, new reservations will continue to accrue. Tesla will phase in Dual Motor and Single Motor so they are satisfied 6 months later. This gives Tesla 2 years to reach economies that will allow it to produce the Single profitably. By this time Tesla will necessarily be producing 500k units a year. However, reservations will continue to accumulate at a rate of 240k a year so it will take a full year+ to work those off. Then Tesla will have more capacity than reservations and will likely phase in Cyber variants that can use the same production lines. However, for the purposes of helping to determine how long someone will have to wait, the ramp during the first 2-3 years is key.

Any hard data to help steer this to be more realistic (if t is not)?

Do you think Tesla will look at the dollars on the table ($50billion+) and ramp up a Cyber factory at an insane speed (faster than China) so as not to loose the reservations?

It is pretty clear that Elon is correct when he says there is enough demand to take 4-5 years to catch up

Update (4/20/20) - methodology is now posted and updated online here:
https://sites.google.com/view/cybertruck-reservation-results/production
Have changed some of the numbers above based on feedback.

Advertisement

 

[TyPope]

I think that Tesla is building 5,000 vehicles a week at their factories which are stuck being set up for three car lines. If you have a factory optimized for a single vehicle, they should be able to pump out many more as long as they can get them to market. I know it's a little long in the tooth, but there is already a factory built in Shreveport, LA that has access to railroad spurs and is set up for vehicle production. It's the old GM Plant where they used to make S-10s and Isuzu Hombres (back when I was there) and more recently, the Hummers.

Now, when I was there, they ran the line at 54 seconds. One vehicle per 54 seconds. That's 533 per shift from one line. That's 1,066 per day working two shifts. That'll get you to 266,666 per year from that one line. Now, it's not perfect but it's already built... there is manpower available... logistics is already built in... the "Line" is already in place. Now, get some industrial engineers in that place and put in the machinery and change it all to fit Tesla. Hey, the building, parking lots, rail yard (there's even a spur INSIDE the building), and a lot of things are already in place. That would make for a quick start!

 

[AlabamaMike]

I think that Tesla is building 5,000 vehicles a week at their factories which are stuck being set up for three car lines. If you have a factory optimized for a single vehicle, they should be able to pump out many more as long as they can get them to market. I know it's a little long in the tooth, but there is already a factory built in Shreveport, LA that has access to railroad spurs and is set up for vehicle production. It's the old GM Plant where they used to make S-10s and Isuzu Hombres (back when I was there) and more recently, the Hummers.

Now, when I was there, they ran the line at 54 seconds. One vehicle per 54 seconds. That's 533 per shift from one line. That's 1,066 per day working two shifts. That'll get you to 266,666 per year from that one line. Now, it's not perfect but it's already built... there is manpower available... logistics is already built in... the "Line" is already in place. Now, get some industrial engineers in that place and put in the machinery and change it all to fit Tesla. Hey, the building, parking lots, rail yard (there's even a spur INSIDE the building), and a lot of things are already in place. That would make for a quick start!

At one time I had seen estimates that a Gigafactory should be able to spit out 500k vehicles a year - assuming multiple lines. Lets hope a Cyber Factory can do such or Tesla will lose almost all their reservations due to not being able to fulfill them. If one believes that is possible - how fast can they rampt to that?

 

[TyPope]

At one time I had seen estimates that a Gigafactory should be able to spit out 500k vehicles a year - assuming multiple lines. Lets hope a Cyber Factory can do such or Tesla will lose almost all their reservations due to not being able to fulfill them. If one believes that is possible - how fast can they rampt to that?

That information was for a single line run on two shifts. Running three shifts, it would be 399,750. Putting in a second line would double all the numbers potentially though I'm not sure there'd be enough room in the factory though it IS 3.1M square feet.

Running 2 lines for 2 shifts would be enough to make 533,332 Cybertrucks per year with time at night for line maintenance/improvements. It would jump start the first production of the Cybertruck while Tesla, perhaps, set up a second factory for more Cybertrucks if needed. The point was that it would be faster than starting at grading the land... maybe. Anyway, here's some info on the site:
GM Shreveport Plant

 

[happy intruder]

I am attempting to add a column to the Cybertruck Reservation tracker that estimates how long someone will have to wait to receive their order after an eventual factory starts production. The tracker is here:

Cybertruck Reservation Tracker

If you scroll up to the top and right in the google sheet you will see the start of assumptions on ramp up. As I worked on estimates, I quickly realized Tesla is in a tight spot. Even assuming rapid ramp up, Tesla is at least 4 years away from catching up. Here are some key assumptions:

• Tesla will start with Tri Motor and move to Dual then Single Motor as economies of scale improve.
• A potentially large number of reservations will cancel out of frustration or may not have the ability to follow through
• Even though the Cybertruck will require considerable new methods of production (along with their own production h*ll), tesla will ramp up quickly: 100k+ within 12 months; 250k+ during second year; 500k+ during year 3 and 4.
• We are 18 months away from the first unit being produced for sale. (6 months to ground breaking; 12 months after that to deliver 1st truck).

What do people think about above? feedback?

As far as reservations:

• reservation rate will quickly hit a steady state of 10% of the US truck market (about 240k a year). Right now the reservation rate is running around 600k a year so this is conservative.
• reservations will continue to build however 25-35% of those will eventually cancel. Does anyone know how this compares to model 3 cancellations?
• Tesla will have to go for it as far as maxing out capability (500k per year) in order to catch up with demand in the next 5+ years; after that they will have to adapt some lines to additional Cyber Vehicles (maybe vans? or cyber cars?) to scale back to a steady state of around 240k per year. If Tesla simply tries to asymptotically ramp to demand it will take 10 years to catch up and people will cancel orders in droves.

What do people think? feedback?

Here is the current state of my uneducated predictions (you can explore the formulas used on the google sheet):

	Tri Motor	Dual Motor	Single Motor	Total	Cumulative				Est Reserv 18m from now
Monthly Rate	25%	30%	35%	<- Additional cancel	288,176	307,969	42,166	638,311	889,061	18 month from now
0-3 month	3,000	0	0	3,000	9,000	0	0	9,000	<- Total Production
3-6 month	6,000	1,000	0	7,000	27,000	3,000	0	30,000
6-9 month	10,000	4,000	0	14,000	57,000	15,000	0	72,000	Total Reserv
9-12 month	15,000	8,000	0	23,000	102,000	39,000	0	141,000	875,561	30 month from now
12-24 month	15,000	15,000	0	30,000	282,000	219,000	0	501,000	1,112,811	42 month from now
24-36 month	16,000	19,000	7,000	42,000	474,000	447,000	84,000	1,005,000	1,350,061	54 month from now
36-48 month	16,000	19,000	7,000	42,000	666,000	675,000	168,000	1,509,000	1,587,311	66 month from now
mix on final year ->	38%	45%	17%		44%	45%	11%	<- cumulative mix

Looking for feedback.

For example we currently estimate over 500k reservations given an assumed cancellation rate and over 600k raw reservations. In 18 months I estimate they will explode to 889k reservations. Between 25% and 35% of those will not follow through (depending on Trim package - see above). This leaves 640k reservations to fill. Tesla will ramp up Tri Motor first to account for low economies of scale and satisfy those Tri Motor reservations within 2 years. However, new reservations will continue to accrue. Tesla will phase in Dual Motor and Single Motor so they are satisfied 6 months later. This gives Tesla 2 years to reach economies that will allow it to produce the Single profitably. By this time Tesla will necessarily be producing 500k units a year. However, reservations will continue to accumulate at a rate of 240k a year so it will take a full year+ to work those off. Then Tesla will have more capacity than reservations and will likely phase in Cyber variants that can use the same production lines. However, for the purposes of helping to determine how long someone will have to wait, the ramp during the first 2-3 years is key.

Any hard data to help steer this to be more realistic (if t is not)?

Do you think Tesla will look at the dollars on the table ($50billion+) and ramp up a Cyber factory at an insane speed (faster than China) so as not to loose the reservations?

It is pretty clear that Elon is correct when he says there is enough demand to take 4-5 years to catch up

well.....on a new production: speed = poor quality so I hope they take their time and document the whole production process and make sure it is repeatable.......

 

[Dixon1430]

That information was for a single line run on two shifts. Running three shifts, it would be 399,750. Putting in a second line would double all the numbers potentially though I'm not sure there'd be enough room in the factory though it IS 3.1M square feet.

Running 2 lines for 2 shifts would be enough to make 533,332 Cybertrucks per year with time at night for line maintenance/improvements. It would jump start the first production of the Cybertruck while Tesla, perhaps, set up a second factory for more Cybertrucks if needed. The point was that it would be faster than starting at grading the land... maybe. Anyway, here's some info on the site:
GM Shreveport Plant

I don't think Elon will ever take over another facility like that again. I could be wrong but what I've read/heard of turning the Fremont factory into an EV production plant was nothing short of a nightmare. The only reason they went with the Fremont factory at that point was out of necessity, they lacked the capital to build a gigafactory and they got a smoking deal from GM. They are continually trying to optimize the plant to function the way they want it to. In the long run I think the building of a new gigafactory would actually be faster to get to production and to ramp up production to any significant numbers.

 

[TyPope]

I don't think Elon will ever take over another facility like that again. I could be wrong but what I've read/heard of turning the Fremont factory into an EV production plant was nothing short of a nightmare. The only reason they went with the Fremont factory at that point was out of necessity, they lacked the capital to build a gigafactory and they got a smoking deal from GM. They are continually trying to optimize the plant to function the way they want it to. In the long run I think the building of a new gigafactory would actually be faster to get to production and to ramp up production to any significant numbers.

From a wholistic approach, you are correct. The Shreveport plant would be low hanging fruit. It would be a way to really shortcut a lot of the time grading a site, permits, etc. But, you are right. They make each facility better than the last one and I like that. With all the logistics already built-in at Shreveport, it MIGHT be worth it to raze the whole place and build the ideal factory there. They have a port as well... Shipping by Truck/Train/Ship is possible there and while they are pretty far South, they ARE in a good spot logistically speaking. Personally, I'd rather they build the plant closer to Nebraska but that's just me.

 

[TyPope]

well.....on a new production: speed = poor quality so I hope they take their time and document the whole production process and make sure it is repeatable.......

I'd posit that each line can pump out one vehicle every 52 to 54 seconds. I don't see how it could go much faster. At this speed, each work station has only those 52-54 seconds to complete their operation. At GM, we would combine a couple of stations (the Cross-Body Harness being one of those) to double the time they had to finish their installations. You had to double the teams to do that. You had one team working on the vehicle and walking along with it for a bit while the 2nd team started on the next vehicle. We also had to have someone dedicated to selecting the wiring harnesses in the correct order (There were 21 different harnesses, a number I thought extremely ridiculous). This is all very possible as it was back then. Tesla seems to be able to automate quite well but parts still have to be delivered to the line at that speed as well. You can figure 8hrs X 3600 seconds per shift which is 533 vehicles... times 5 days is 2,666 per week or 133,333 per year (50 weeks of production and 2 weeks for maintenance/holiday).

Looked at another way, each line that outputs a vehicle has approximately 30,326,400 seconds of operation time possible. (That's 24hrs x 5 days x 50 weeks). 30,326,400 seconds / 500,000 vehicles = 1 vehicle every 60.65 seconds pretty much every hour of every day. For reference, Tesla made 6,000 vehicles per week which is one every 100.8 seconds when they were pumping out the Model 3 like mad.

Basically, to realistically be able to make 500,000 Cybertrucks a year, they need 2 lines running almost 24/7. And I haven't even touched on the logistics of moving 1,424 Cybertrucks every single day... That's a lot of carriers (6 Cybertrucks per trailer = 237 trucks per day... One every 6 minutes.


Whew... sorry about the word vomit in this post.

 

[AlabamaMike]

Basically, to realistically be able to make 500,000 Cybertrucks a year, they need 2 lines running almost 24/7. And I haven't even touched on the logistics of moving 1,424 Cybertrucks every single day... That's a lot of carriers (6 Cybertrucks per trailer = 237 trucks per day... One every 6 minutes.

Tesla appears to take continuous process improvement to heart. Each Gigafactory appears to be more aggressive in scope and time. There are articles from Germany claiming that the G4 is being designed to output 500,000 cars a year. Not sure if that is wishful thinking or comes from Tesla. I used that in my assumptions above. China demonstrated 3000 per week within 3 months and is expected to ramp to 5000 per week by only adding shifts. I am sure there will be lots of logistical issues with 150 car carriers per day. However, this is also with at most 1/2 of a GF built out. I would not bet against Tesla eventually hitting 500k cars per year from GF3. The question is how fast can this be done and what does the ramp up look like. Maybe it comes in spurts as lines or shifts are added. However, other logistcal nightmares might lead to more linear ramp?

On a side note - if you contributed to the Cybertruck Reservation Tracker - check back and your reservation should now have an estimated delivery time. This is obviously subject to revision as the ramp up estimate is refined.

 

[TyPope]

Tesla appears to take continuous process improvement to heart. Each Gigafactory appears to be more aggressive in scope and time. There are articles from Germany claiming that the G4 is being designed to output 500,000 cars a year. Not sure if that is wishful thinking or comes from Tesla. I used that in my assumptions above. China demonstrated 3000 per week within 3 months and is expected to ramp to 5000 per week by only adding shifts. I am sure there will be lots of logistical issues with 150 car carriers per day. However, this is also with at most 1/2 of a GF built out. I would not bet against Tesla eventually hitting 500k cars per year from GF3. The question is how fast can this be done and what does the ramp up look like. Maybe it comes in spurts as lines or shifts are added. However, other logistcal nightmares might lead to more linear ramp?

On a side note - if you contributed to the Cybertruck Reservation Tracker - check back and your reservation should now have an estimated delivery time. This is obviously subject to revision as the ramp up estimate is refined.

I did check back. Nice feature add.

 

[AlabamaMike]

Started a web site to post results from the Tracker including delivery estimates. This is so the results can be presented in the proper context.

https://sites.google.com/view/cybertruck-reservation-results/home

 

[markymark]

My two cents worth... I think these estimates are not conservative enough, I think it is better to approach this "semi"-modeling exercise (pun intended) conservatively rather than tending towards overestimating - if the GF3 happens sooner, if the CT is pumped out quicker on the line, if less people cancel, if more people reserve - that's all good. The Y is largely based on the 3, CT is a whole new beast altogether - materials, processes, machining, etc., I do not think you can use Fremont/GF1/GF2 experiences to neatly map onto GF3 or the CT production cycle. Like the Model Y, I think setting conservative expectations and then exceeding them by delivering early is way better all around for everyone. If that means losing some CT reservations, I don't see that as an issue for Tesla, as demand will outstrip supply for some time. I also think the number of reservations in inflated - I know I have three right now, one of each flavor because I don't know what my financial situation will be and/or how bad my desire to get one right away will override that!

 

[Dixon1430]

My two cents worth... I think these estimates are not conservative enough, I think it is better to approach this "semi"-modeling exercise (pun intended) conservatively rather than tending towards overestimating - if the GF3 happens sooner, if the CT is pumped out quicker on the line, if less people cancel, if more people reserve - that's all good. The Y is largely based on the 3, CT is a whole new beast altogether - materials, processes, machining, etc., I do not think you can use Fremont/GF1/GF2 experiences to neatly map onto GF3 or the CT production cycle. Like the Model Y, I think setting conservative expectations and then exceeding them by delivering early is way better all around for everyone. If that means losing some CT reservations, I don't see that as an issue for Tesla, as demand will outstrip supply for some time. I also think the number of reservations in inflated - I know I have three right now, one of each flavor because I don't know what my financial situation will be and/or how bad my desire to get one right away will override that!

I think you have a great point to a degree. I think that you could speculate that the ramp of the cybertruck will be in between the model 3 ramp and the model y ramp. I agree that I don't think that tesla will be able to ramp the CT the same way they have with the Y but I also think that everything they have learned along the way will help smooth out some of the unknowns that they had to deal with when ramping up model 3 production.

I guess all this to say, I think it will be much faster/smoother than model 3 ramp, just not as fast and as smooth as the Y.

Hopefully we are both wrong and Elon blows us away once again.

 

[TechOps]

I think you have a great point to a degree. I think that you could speculate that the ramp of the cybertruck will be in between the model 3 ramp and the model y ramp. I agree that I don't think that tesla will be able to ramp the CT the same way they have with the Y but I also think that everything they have learned along the way will help smooth out some of the unknowns that they had to deal with when ramping up model 3 production.

I guess all this to say, I think it will be much faster/smoother than model 3 ramp, just not as fast and as smooth as the Y.

Hopefully we are both wrong and Elon blows us away once again.

I agree; my estimate is that 20-25% of current reservations are "legit". The other 75-80% of them will fall away. I made two reservations on launch night, and initially hated the truck but it's grown on me so much that I'll be buying one. My other reservation will xfer to a family member or be canceled. I imagine there are many people in the same boat, especially for reservations made in the first week.

 

[DMC-81]

Of course, after listening to the last earnings call, the governor on the ramp and speed estimates for the CT will be the rate of battery production.

I'm looking for good news on battery day!