Gigafactory locations and products

[Buckminster]

size comparison.... wonder how large Giga Texas will end up being

Fairy Modfather: edited to show the graphic, otherwise no point. --ggr

 

[Mo City]

Fairy Modfather: edited to show the graphic, otherwise no point. --ggr

Viewing the Giga Texas area on Google Maps makes me even more excited about the project. It think it's really cool so much of the eastern boundary is the Colorado River.

The main things that bother me are the property being bisected by Hwy 130 and the power lines. Can the power lines be moved?

 

[MC3OZ]

This is some speculation on what Tesla growing at Max-Q might looking like.
You can consider my version of the plan a "worked example" that may not necessarily be much like Tesla's actual plan.
The reason for this "deep dive" is consider the constraints on growth real and imagined, and consider how Tesla might overcome some constraints to maximize growth and accelerate the mission.

My starting point is estimates posted by Gary Black on twitter.
Fremont - Current 490K - 2H 2020 + 100k
Shanghai - Current (3) 200K - 2H 2020 (Y) + 200K
Berlin - 2021 - Y+3 +400K
Austin - 2021 Y + 200K - 2021 Cybertruck +200K

Total production capacity by year end 2021 approx 1.8M vehicles per year...

Considering the limitations on future growth.

Not an issue:-

• Capital
• Demand

Can be solved with money:-

• Land
• Buildings
• Production machinery.

More difficult:-

• Raw materials
• Qualified staff

Tesla is nearly always on time with problems that are merely difficult, especially if it is something they have done before.
New factories in new locations, to build existing vehicle designs, are essentially a cut-and-paste..

Battery Day should reveal that production of energy storage batteries via Roadrunner lines in new factories, is not a difficult problem.

For some new factories Energy Storage Batteries may be how they start out, IMO likely candidates are:-

• UK
• Australia
• India

Which raises the question of "Qualified staff" - staff from the US, Germany or China can help with initial production and training of local staff. Having an established factory helps train additional staff.

If these locations progress to vehicle production, local staff can be partially trained by working in US, German or Chinese factories and again US, German and Chinese staff can help with the initial factory set up and training of local staff.

I don't want to make this post too long, but at first these locations progressing to vehicle production seems surprising, but what is required is making the right type of vehicles in the right locations, in the right volumes at the right time.

For example, in Australia a smaller Cybertruck is the ideal starting vehicle in a production volume of 50K per year.
In fact I suggest the US design team focuses mainly of Cyber design, and the German and Chinese teams focus on compact designs.
The reason for this is Cyber designs can be built in smaller volumes hence built in more locations, they are an ideal way to start factories in new countries.

Having a factory in a local country, state, or city is a great marketing tool.

Another way new factory locations can start is by doing GA and paint correction on imported Model 3/Y, and perhaps later other models.

When we look at expansion by acquisitions and partnerships there are broad categories that possibly make sense:-

• Mining and processing
• Supplying Vehicle drive-trains and components, (e.g supplying Indian car makers)
• Joint vehicle programs - e.g. British sports car with Lotus or Aston Martin.
• Acquiring other types of vehicle production - Buses, Garbage Trucks, Farm Machinery, Earth Moving

Without turning this post into a novel, I've barely scratched the surface of the potential expansion opportunities, and handled correctly, with smart and pragmatic decision making, most perceived problems can be overcome..

So IMO the bottom line is many more factories, models and vehicle types no reason why Tesla can't make 20M vehicles per year, sooner than most people anticipate.

 

[StealthP3D]

Considering the limitations on future growth.

Not an issue:-

• Capital
• Demand

Can be solved with money:-

• Land
• Buildings
• Production machinery.

You make a common mistake of thinking money is the primary limitation on acquiring suitable land, buildings and production equipment and that these issues can be "solved" with money. The reality is that a lot of expertise is required to acquire land, buildings and production machinery that leads to success. It also needs to be done in a capital efficient way. Something as basic as choosing the factory location can be the difference between success and failure and the decision is not straight-forward or simple - it requires a lot of research, knowledge and insight. Even then it's like picking a winning stock. You had better be good at it. All of these things take massive amounts of time and expertise. So, even these simple things that you portray as being solved with money requires talented people who are "up to speed" on the overall game plan and what is required.

 

[Buckminster]

I think that another limitation on expansion is Elon wanting to improve the machine that builds the machine with each iteration. He could create more Shanghai's - cookie cutter style but Berlin is a step forward which takes significant engineering resource.

Elon follows marginal gains methodology except each step is a bit more than marginal.

 

[MC3OZ]

You make a common mistake of thinking money is the primary limitation on acquiring suitable land, buildings and production equipment and that these issues can be "solved" with money. The reality is that a lot of expertise is required to acquire land, buildings and production machinery that leads to success. It also needs to be done in a capital efficient way. Something as basic as choosing the factory location can be the difference between success and failure and the decision is not straight-forward or simple - it requires a lot of research, knowledge and insight. Even then it's like picking a winning stock. You had better be good at it. All of these things take massive amounts of time and expertise. So, even these simple things that you portray as being solved with money requires talented people who are "up to speed" on the overall game plan and what is required.

Fair points, but mainly for the larger more important factories.

For factories that are smaller and less important, location is not so critical, but being in the same country as the end customer an the raw materials supply chains is a big plus. Still mostly what is required is time and expertise, some of that expertise has to be local.. But in most cases 6-24 months should be sufficient to find the right location, if the requirements are well defined.

I also think we may find moving a Roadrunner style battery line might not be a big deal.. yes it probably will cost if they get it wrong... but perhaps not a significant amount.

For a smaller scale lower capex production process that are essentially replicating an existing process, the machinery to use is exactly the same machinery set up in exactly the same way.

Expertise is important when you haven't found the formula, the US, China and Germany are places where Tesla should create and refine the formula, in other places they should get apply the formula unaltered. Once they build expertise in other locations they can contribute more.. but when you want to move quickly an seize opportunities while they are there, simplicity and sticking to a winning formula is important..

 

[CorneliusXX]

This is some speculation on what Tesla growing at Max-Q might looking like.
You can consider my version of the plan a "worked example" that may not necessarily be much like Tesla's actual plan.
The reason for this "deep dive" is consider the constraints on growth real and imagined, and consider how Tesla might overcome some constraints to maximize growth and accelerate the mission.

My starting point is estimates posted by Gary Black on twitter.
Fremont - Current 490K - 2H 2020 + 100k
Shanghai - Current (3) 200K - 2H 2020 (Y) + 200K
Berlin - 2021 - Y+3 +400K
Austin - 2021 Y + 200K - 2021 Cybertruck +200K

Total production capacity by year end 2021 approx 1.8M vehicles per year...

Considering the limitations on future growth.

Not an issue:-

• Capital
• Demand

Can be solved with money:-

• Land
• Buildings
• Production machinery.

More difficult:-

• Raw materials
• Qualified staff

Tesla is nearly always on time with problems that are merely difficult, especially if it is something they have done before.
New factories in new locations, to build existing vehicle designs, are essentially a cut-and-paste..

Battery Day should reveal that production of energy storage batteries via Roadrunner lines in new factories, is not a difficult problem.

For some new factories Energy Storage Batteries may be how they start out, IMO likely candidates are:-

• UK
• Australia
• India

Which raises the question of "Qualified staff" - staff from the US, Germany or China can help with initial production and training of local staff. Having an established factory helps train additional staff.

If these locations progress to vehicle production, local staff can be partially trained by working in US, German or Chinese factories and again US, German and Chinese staff can help with the initial factory set up and training of local staff.

I don't want to make this post too long, but at first these locations progressing to vehicle production seems surprising, but what is required is making the right type of vehicles in the right locations, in the right volumes at the right time.

For example, in Australia a smaller Cybertruck is the ideal starting vehicle in a production volume of 50K per year.
In fact I suggest the US design team focuses mainly of Cyber design, and the German and Chinese teams focus on compact designs.
The reason for this is Cyber designs can be built in smaller volumes hence built in more locations, they are an ideal way to start factories in new countries.

Having a factory in a local country, state, or city is a great marketing tool.

Another way new factory locations can start is by doing GA and paint correction on imported Model 3/Y, and perhaps later other models.

When we look at expansion by acquisitions and partnerships there are broad categories that possibly make sense:-

• Mining and processing
• Supplying Vehicle drive-trains and components, (e.g supplying Indian car makers)
• Joint vehicle programs - e.g. British sports car with Lotus or Aston Martin.
• Acquiring other types of vehicle production - Buses, Garbage Trucks, Farm Machinery, Earth Moving

Without turning this post into a novel, I've barely scratched the surface of the potential expansion opportunities, and handled correctly, with smart and pragmatic decision making, most perceived problems can be overcome..

So IMO the bottom line is many more factories, models and vehicle types no reason why Tesla can't make 20M vehicles per year, sooner than most people anticipate.

I think many of these niche products and small population geographies are unlikely to be targeted by Tesla (Unless Tesla can think of a unique value add) for quite a long time.

Judging by Elon's comments on manufacturing efficiency (treating the factory like a microprocessor and increasing the volumetric density and clock speed) I think it lends itself to fewer but larger (by production volume) factories. Using rough numbers, Tesla can make around 2,500 vehicles per week on a single line in Fremont. If that speed increases 10x as he is planning then the target production in Australia would be met with 2 weeks of work - leaving the line idle for the rest of the year. If Tesla produce vehicles at a slower pace than they can there are huge overheads in the cost from additional fixed cost allocation and higher variable costs (e.g wages). It wouldn't take much for these additional overheads to be more costly than just shipping the vehicles from the high volume production factory which has lower cost overheads. Cell production for stationary storage may work out as this appears to be a highly parallelised process - potentially leading to smaller factories still being able to run at peak efficiency.

I grew up on a farm in Australia and in terms of farming equipment, I don't think they lend themselves to EVs with the current technology for a few reasons:

• Large tractors tend to be left in the shed 11 months of the year then run flat out (20+ hours a day) as soon as conditions are right for ploughing/sowing (e.g. after rainfall). There's no pack that could allow tractors to do that work as recharging would take too long.
• A pack large enough to run a tractor would likely make it too heavy and end up getting bogged.
• The cost of a tractor with a pack that large would be prohibitively expensive (for now)
• The same would apply to harvesters which again run flat out during harvest season for 20 hours a day. If there's any slowdown in harvesting and rain falls on a ripe crop it is devalued substantially and potentially dangerous to store. So not worth the risk at the moment.

Smaller tractors that are only used for odd jobs or spraying might be easier to convert as they don't have the time pressure of the above equipment. A modified cybertruck might even be able to replace some of that work with its towing capacity and low speed power.

My guess is that earth moving equipment is likely to have similar time pressures where the recharging speed would be the limiting factor in enabling the transition. Buses, delivery and garbage trucks would likely be more flexible as you could design routes to fit the operational constraints of the vehicles - just like what will be done with the Semi.

 

[Christine69420]

I grew up on a farm in Australia and in terms of farming equipment, I don't think they lend themselves to EVs with the current technology for a few reasons:

I don't know anything about farming so my comments are general and not for that particular use case.

I know that in Norway they have developed an electric excavator. It weighs 26 tons and can work 5-7 hours between each charging session.

Source: Google Translate

My thoughts is that ie Caterpillar could standardize on a particular battery dimension to be used in all of it's machines when they go electric. Then on building sites they could have enough batteries charging so that machines could swap used batteries for fully charged ones when needed.

Much like we do today with electric power tools.

As for weight - how much do a diesel engine + fuel tank weigh vs batteries and an electric motor?

In addition - some projects in Norway already demand the use of electric machines to keep noise levels down for the neighbors of the site. Which means that the contractor can send the extra cost - if any - to the paying customer.

 

[Zenn]

I think many of these niche products and small population geographies are unlikely to be targeted by Tesla (Unless Tesla can think of a unique value add) for quite a long time.

Judging by Elon's comments on manufacturing efficiency (treating the factory like a microprocessor and increasing the volumetric density and clock speed) I think it lends itself to fewer but larger (by production volume) factories. Using rough numbers, Tesla can make around 2,500 vehicles per week on a single line in Fremont. If that speed increases 10x as he is planning then the target production in Australia would be met with 2 weeks of work - leaving the line idle for the rest of the year. If Tesla produce vehicles at a slower pace than they can there are huge overheads in the cost from additional fixed cost allocation and higher variable costs (e.g wages). It wouldn't take much for these additional overheads to be more costly than just shipping the vehicles from the high volume production factory which has lower cost overheads. Cell production for stationary storage may work out as this appears to be a highly parallelised process - potentially leading to smaller factories still being able to run at peak efficiency.

I grew up on a farm in Australia and in terms of farming equipment, I don't think they lend themselves to EVs with the current technology for a few reasons:

• Large tractors tend to be left in the shed 11 months of the year then run flat out (20+ hours a day) as soon as conditions are right for ploughing/sowing (e.g. after rainfall). There's no pack that could allow tractors to do that work as recharging would take too long.
• A pack large enough to run a tractor would likely make it too heavy and end up getting bogged.
• The cost of a tractor with a pack that large would be prohibitively expensive (for now)
• The same would apply to harvesters which again run flat out during harvest season for 20 hours a day. If there's any slowdown in harvesting and rain falls on a ripe crop it is devalued substantially and potentially dangerous to store. So not worth the risk at the moment.

Smaller tractors that are only used for odd jobs or spraying might be easier to convert as they don't have the time pressure of the above equipment. A modified cybertruck might even be able to replace some of that work with its towing capacity and low speed power.

My guess is that earth moving equipment is likely to have similar time pressures where the recharging speed would be the limiting factor in enabling the transition. Buses, delivery and garbage trucks would likely be more flexible as you could design routes to fit the operational constraints of the vehicles - just like what will be done with the Semi.

Given that a harvester uses approximately 13.5 gallons of diesel fuel per hour with an efficiency of 20%, you would need about10 kwh in battery power every hour or 200 kWh for 20 hour day.
Converting Inefficient hydraulic systems into electric drives would lessen energy requirements.
You would need to keep your equipment plugged-in when not in use especially in cold weather.
If the above is right?...the use of electric power for harvesting is feasible but would be costly, as you noted.

 

[CorneliusXX]

Given that a harvester uses approximately 13.5 gallons of diesel fuel per hour with an efficiency of 20%, you would need about10 kwh in battery power every hour or 200 kWh for 20 hour day.
Converting Inefficient hydraulic systems into electric drives would lessen energy requirements.
You would need to keep your equipment plugged-in when not in use especially in cold weather.
If the above is right?...the use of electric power for harvesting is feasible but would be costly, as you noted.

Are you sure that works out? I can't imagine a Semi would need 4-5x the battery capacity of a harvester running 20 hours to go 500 miles. The roadster is expected to have a battery pack of 200KWh which seems incredibly small to run a harvester all day.

 

[MC3OZ]

I don't know anything about farming so my comments are general and not for that particular use case.

I know that in Norway they have developed an electric excavator. It weighs 26 tons and can work 5-7 hours between each charging session.

Source: Google Translate

My thoughts is that ie Caterpillar could standardize on a particular battery dimension to be used in all of it's machines when they go electric. Then on building sites they could have enough batteries charging so that machines could swap used batteries for fully charged ones when needed.

Much like we do today with electric power tools.

As for weight - how much do a diesel engine + fuel tank weigh vs batteries and an electric motor?

In addition - some projects in Norway already demand the use of electric machines to keep noise levels down for the neighbors of the site. Which means that the contractor can send the extra cost - if any - to the paying customer.

This video shows one possible solution:-

If we imagine a farm machine with a built in 50 KWh battery, extended by 20 x 7 KWh battery bricks.
The battery bricks are like lego bricks, all that is needed is a standard connection protocol, size and weight...

This machine would have 190 KWh, but 140 kWh could be replaced in 5 minutes buy swapping bricks, the built in battery allows bricks to be fully discharged and for the swap to happen at a convenient time.

Say the farmer owns 40 bricks which he charges with his own solar, at harvest time he can hire another 100 bricks which arrive fully changed.

The hire company can also provide charged bricks to other farmers, construction companies and households...

The farm machine can have a 3kW solar canopy which proves another 20 kWh per day on a sunny day.

In terms of Tesla being involved, only 3 questions are relevant:-

1. Is this part of the mission?
2. Do they have good solution, which will be attractive to customers?
3. Can they make money?

If Tesla decides the answer to 1. is Yes, they can find a way to solve 2. & 3.

 

[Buckminster]

Freemont (not strictly a Gigafactory)
Minimal battery build
M3/MY - 2170 (I'm assuming no change in short term)
MS/MX - 18650 (I'm assuming no change in short term)
Roadster

Gigafactory Nevada (previously GF1)
2170s (10% improved - new chemistry only following Pana deal)
4070 cells from 2020/21
Powerwalls - 2170s from Nevada then 4070s
Powerpacks - 2170s from Nevada then 4070s
Megapacks - 2170s from Nevada then 4070s
Power trains for Freemont
Possibly Semi from 2021?
Nevada seems to be out of favour.

Gigafactory New York (previously GF2)
Solar panels
Solar roof

Gigafactory Shanghai (previously GF3)
2170s
M3/MY - 2170 + Iron Phosphate (supplied) (I'm assuming no change in short term)
I see little change in the above for 3 years.

Gigafactory Berlin (previously GF4)
2170s (outsourced)
M3/MY - 2170 (I'm assuming no change in short term)
I see little change in the above for 3 years.

Gigafactory Texas in Austin
Will be producing 4070 cells only
Cybertruck from 2021 with 4070s (tent initially?)
Semi from 2021 with 4070s
MY (using 2170s from Nevada) from 2022
M3?

Gigafactory Bristol/London?
2170s (outsourced)
M3/MY - 2170 (I'm assuming no change in short term)
I see little change in the above for 3 years.

Gigafactory Guangdong?
4070 cells only
Semi from 2022
M2 from 2023

NEW Gigafactory Pennsylvania? (North East) Breaks ground 2024
4070 cells only
Semi? from 2025
M2 from 2025

Gigafactory Seoul?

 

[MC3OZ]

M3/MY - 2170 (I'm assuming no change in short term)
MS/MX - 18650 (I'm assuming no change in short term)

I'm not sure I would make these assumptions as the Kato Rd line is making cells for something, and we will find out what on battery day.

Plaid Model S is a likely contender.... but I would not rule anything in or out ..

I do think more factories are possible/likely once Berlin and Texas are close to fully built..

An additional Chinese factory and another factory in Europe and or the UK seem like the next logical step.

 

[Zenn]

Are you sure that works out? I can't imagine a Semi would need 4-5x the battery capacity of a harvester running 20 hours to go 500 miles. The roadster is expected to have a battery pack of 200KWh which seems incredibly small to run a harvester all day.

No...I’m not sure if it would work out as I calculated but it should not be too far out given the fuel consumption of today's harvesters is accurate. But I do believe it is feasible and likely practical...it needs to be demonstrated.
However, a semi weighs 80,000 lbs and has perhaps 14 or more wheels generating rolling friction plus significant air drag at highway speeds. I believe a semi requires approximately 2 KWh per mile or 2 kWh per minute at 60 miles per hour. The low speeds of farm equipment would use a fraction of a semi's energy requirement. Then you need to add the energy needed to run farming implements.

 

[petit_bateau]

From the main thread:

Has anyone attempted to model out an estimate of cash needs for capex over the next few years building up on a per project basis rather than a % of either current capex or some other ratio? The reason I ask it that it feels like there needs to be a pretty rapid expansion in capex spending to meet all the upcoming expansion plans. I think we've had it a bit easy with the Chinese expansion as they've just been handing out free money so it's probably not a good baseline assumption for cash needs on upcoming expansion plans.
===========
I have done this in my modelling. It seems to me that one needs to split out the site level, and the line level. My outcome was that at some point during 2020 TSLA are no longer capex-constrained, and I think that within a year or so they will not be human-resource constrained (the experienced mge cadre is what it takes to build/operate multiple sites simultaneously). The main issue in this respect seems to me to be how fast they can build 'sets' of production lines, and it is likely that the constraints are the cell lines and/or the large casting machines. Probably it is the cell lines that are the more durable constraint.

(please excuse me, but I'm not very practiced at using the quote function on this BB)

regards, dspp

 

[Christine69420]

From Teslamag.de:

8 Giga presses in Gigafactory Berlin

Tesla boss announces radical new technology for production in Giga Berlin: This is what it looks like:

One of the new Giga presses was recently installed in Fremont and in Tesla's second new Gigafactory building in China . And for Tesla's German Gigfactory in Grünheide, Brandenburg, eight of these hundred-ton machines are planned.

"There are 8 die-casting machines with a maximum operating capacity of 73.1 t / d per machine", says one of the thousands of pages of Tesla applications for the German Gigafactory , which were officially laid out on the Internet and documented by a Twitter user .

Source: Tesla-Chef: Transformation in Giga Berlin – Pläne zeigen sie > teslamag.de

 

[RubberToe]

Can someone who speaks German please translate these 4 slides about rail connections to the Berlin Gigafactory?

https://twitter.com/Gf4Tesla/status/1302691861807747072/photo/1

 

[mblakele]

Can someone who speaks German please translate these 4 slides about rail connections to the Berlin Gigafactory?

https://twitter.com/Gf4Tesla/status/1302691861807747072/photo/1

Hmm — I thought I could read German, but these slides are full of technical vocabulary that I don't know. We have a few native speakers of German on TMC though. For example maybe @avoigt could help?

Is there specific information you're looking for? From what I can make out, twitter's GF4Tesla summarized it well:

These plans take into account local public transport and the supply of materials and removal of finished cars.​

 

[Krash]

Can someone who speaks German please translate these 4 slides about rail connections to the Berlin Gigafactory?

https://twitter.com/Gf4Tesla/status/1302691861807747072/photo/1

Convertio meets reverso attached. Don't shoot the messenger.






Below was another pairing of OCR meets translation in just text...

Fangschleuse station Connection tesla plant to rail infrastructure Current status

Description of the existing infrastructure fang lock station

Legend: — Existing building — New construction --• New construction (option) Dismantling Of IAV limit

Fangschleuse station contains essential infrastructural elements: two outer platforms at the BÜ for use by the SPNV line RE1 (useful length = 135 m) - two overtaking tracks (useful lengths = approx. 660 and 670 m respectively) - Connection of the DRE to the GVZ Berlin Ost Freienbrink

Next steps DB NETZE

Various activities are currently underway

• Coordination of the target infrastructure with Tesla and integration into the higher-level planning (e.g. development plan)

Consideration of a new location of the Fangschleuse transport station with the participation of DB Station&Service, Land Brandenburg and Tesla - Investigations into the feasibility of SPV in the plant with the participation of Tesla and VBB - Integration into the overall development concept (free space, integration into the road network, parking spaces, interchanges)

• Votes on project scope and funding

Previous approach of DB Netz AG Development of a preferred variant with capacity assessment

— Existing buildings — new construction --• New construction (option) Dismantling of IAV limit

Batretelgzufollgung (Tunnel Box)

DB Netz AG has developed a preferred variant for the target infrastructure: Taking into account all the above-mentioned requirements - variant has been roughly routed

DB Netz AG has examined the capacity of the route and the preferred variant: Definition of an additional approx. 24 pairs of pulls from/to Fangschleuse/Tesla as upper limit . Higher values require additional adjustment requirements in fishing lock or other areas of the net (e.g. Wuhlheider Kreuz)

 

[avoigt]

Hmm — I thought I could read German, but these slides are full of technical vocabulary that I don't know. We have a few native speakers of German on TMC though. For example maybe @avoigt could help?

Is there specific information you're looking for? From what I can make out, twitter's GF4Tesla summarized it well:

These plans take into account local public transport and the supply of materials and removal of finished cars.​

I understand all of it. What is your questions?