Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

This site may earn commission on affiliate links.
Two recent Tesla news items just fell into place for me (after a bit of delay) after watching Rob´s video from yesterday:
- A rumour about new version of 4680 production machinery
- The permit application for a new Battery equipment production line to be built at Fremont

I didn´t click for me at the time, but looking at it now, it seems obvious that if the rumour is true, they need to make a series of these production machines as fast as possible for Austin and Berlin so that´s what they´d to with the new line in Fremont IMHO.
This SMR video is additional context:-

Elon is focusing on solving the problem of scaling battery raw materials, IMO the reflects some confidence that they can solve any issues with 4680 production.

The think about what Tesla can do with volume 4680 production, structural battery packs and Giga-castings. That is the perfect formula for ramping volume production in a capex efficient way, constrained by equipment availability, workforce, construction times, and raw materials.

Multiple factories is a partial solution to workforce and construction times, but ultimately factories can't run ahead of the supply of raw materials,.
 
No, volume at the Closing Cross yesterday was > 8M shares:

View attachment 847970

Today's Closing Cross volume was < 1.2M shares:

View attachment 847971

Yet total volume for these 2 days was within 1.2M shares of eachother, with today having the higher volume.

TL;dr major manipulation by shortzes during the day yesterday, likely most of which was covered at the Closing Cross.
Thanks Artful
 
Taiwan's August new car sale figure is out. As usual, usually first month (July) of the quarter sees almost no sales but starts to jump and takes lead in second and 3rd month because they receive shipments from Fremont factory.

1. Toyota Corolla Cross 3216 (hybrid 903)
2. Tesla Model 3 1774
3. Toyota Altis 1321 (hybrid 328)
4. Nissan Kicks 1005
5. Honda CR-V 932
6.. Ford Focus 896 (Active 410)
7. Ford Kuga 860
8. Toyota RAV4 770
9. Honda HR-V 760
10. Toyota Yaris 695
11. Volkswagen Tiguan 693
12. Toyota Sienta 581
13. Mazda CX-5 464
14. Hyundai Venue 454
15. Nissan Sentra 438
16. Toyota Vios 427
17. Toyota Camry 385
18. Volkswagen Golf 371
19. Honda Fit 358
20. M. Benz GLE 351

Source(traditional Chinese): 民俗月讓國產車「度小月」 8 月新車銷售 Tesla 與 VW 交車量大爆發 - 自由電子報汽車頻道
 
Cathie doesn't look at earnings, forward PE, PEGs, or anything on the P&L. She only cares about stories of the stock.
She invents the stories of her 5 pillars then pretends companies fill them. I've posted on here, probably two full years ago, about looking into companies in her portfolio and finding them to be complete junk. Only one I actually liked (and bought), Coinbase, and we know how that turned out.
 
Tesla, Toyota, and The Next Generation of Lean Manufacturing
The Toyota Production System and the Toyota Way revolutionized not only the automotive industry but also the entire global manufacturing sector. I've worked as a quality engineer at Boeing and another major industrial company in a different industry. Like many companies, my employers designed their own production systems basically as rebranded variants of the TPS adapted for their own purposes. In manufacturing, it is common to use Japanese words taken straight from the TPS, like andon, poka-yoke, kanban, gemba, kaizen, muri, and muda. The propulsion shop at my Boeing plant even had posters on the wall of Taiichi Ohno and some of his most famous sayings. The prosperity we live in today is in no small part due to the contributions of the Toyota Production System and the Toyota Way to changing the way businesses are run.

The fact that Tesla is now blowing the TPS out of the water is, therefore, astonishing. However, most of what Tesla is doing in their manufacturing efforts is not really new fundamentally (and the same goes for SpaceX and Boring). Tesla's production system is based on the same principles established in the TPS and Toyota Way and raising the standard of execution on them in practice to new levels never seen before, while massively amplifying the effect with modern computer technology. I don't know if Tesla is intentionally copying Toyota and pulverizing them at their own game, but in effect that is what's happening. There has been no sign of Tesla using Toyota's terminology and they've never said they're learning from Toyota, so it's quite possible Elon has reinvented the wheel on a lot of this by failing a bunch and learning quickly from his mistakes, but in the end it doesn't really matter why he follows the principles because either way the principles work and the outcome is the same.


Elon came up with some catchier phrasing for the principles but the foundations of Lean Manufacturing have been known for longer than he's been alive. For instance, compare Toyota's "8 Wastes" to principles Tesla uses for industrial engineering.
  1. Waste of overproduction (largest waste)
  2. Waste of time on hand (waiting)
  3. Waste of transportation
  4. Waste of processing itself
  5. Waste of excess inventory
  6. Waste of movement
  7. Waste of making defective products
  8. Waste of underutilized workers
Reducing these forms of waste is the first principle of industrial engineering. This is self-evident. If you expend valuable resources doing any these 8 things that the customer doesn't value (or values negatively), then by definition you are wasting resources. An effective organization transforms inputs of a certain value into outputs of a higher value.

The key problem with adopting Lean and attempting to reduce waste is not in understanding the framework, but rather the problem is, and always has been, psychological factors, especially within management and even more so in top management. Ego, fear, lack of teamwork, arrogance, laziness, greed, distrust, resistance to change, short-termism, and other age-old human flaws stand in the way of the right stuff happening. I believe that this is fundamentally where Tesla and Elon Musk excel, and why they are now stomping on Toyota. Manufacturing will be Tesla's long-term competitive advantage and in this essay I want to summarize my thesis for why.

After WW2, American armed forces and other Allied powers continued occupying Japan, helping rebuild Japanese society and industry, in part due to a desire to increase American influence in East Asia and prevent Japan from descending into communism under Soviet shadow control, as eventually happened in North Korea, China, Mongolia, Vietnam and many other Southeast Asian nations. In 1947, General Douglas MacArthur requested that a statistician named Dr. William Edwards Deming be transferred to Japan to help the government census in Japan and to help with rebuilding Japan in general. Deming had an academic background in physics and mathematics at Yale and prior to his first trip to Japan he had been working for the US Census Bureau helping them implement much better internal quality processes and statistical sampling techniques.

Twenty years prior, Deming had learned radical new techniques and management frameworks while working at Bell Labs with Walter Shewhart, a fellow physicist who had pioneered modern statistical process control. As physicists who came of age in an era during which the very philosophical foundations of science were being questioned question as theories like Relativity and Quantum Mechanics were coming into vogue, both men had a keen intellectual interest in epistemology, the study of where knowledge comes from. As scientists by training, both also possessed the skills and mathematical understand to actually apply the Scientific Method correctly. They transferred this knowledge and mindset into their ideas for improving organizational processes with rigorous empirical techniques. Deming was enamored with Shewhart's ideas and helped Shewhart develop them further. By the time the war happened, Deming had gained a substantial reputation, especially within the federal government after having improved efficiency at several departments and especially during the war when he trained many American workers on quality improvement.

At the time, Japan had a well-deserved reputation for producing junk. They had serious problems with quality and low productivity, and so the Japanese Union of Scientists and Engineers had keen interest in improving. Meanwhile, North America was the only area of the world with advanced industrial capacity that hadn't been damaged by the war, and America in particular had undergone a revolution in quality management and statistical process control in the years leading up to WWII and then especially during the war when production was mission-critical and they had to figure out how to quickly get women trained and up to speed on the production lines. Many of the techniques of acceptance sampling and statistical process control developed during this period are still in use today, and I personally used them while at Boeing. Deming himself was quite influential in the spread of these techniques for supporting the war effort. The Japanese, humbled after losing the war, were ready to learn from what others, particularly the Americans, were doing so much better than they were. Meanwhile, the Americans took a turn away from what had helped make the war production effort so successful, and they began focusing less on quality and more on piece cost reduction and cranking out maximum production quantity at the expense of everything else. This appeared to work for a while, largely because of the USA having overwhelmingly the best postwar industrial capacity which fed the immense demand from American allies whose factories and roads had been leveled by bombs, and also because of the cheap domestic American oil and coal of the 20th century. Japanese engineers and industrial leaders frequently visited America during this time and absorbed knowledge from industrial titans like Ford, Bell, and Westinghouse, and even American retail and grocery stores.

So, while working for the US government rebuilding efforts in Japan, in 1950 Deming received requests from the Japanese Union of Scientists and Engineers to give lectures and hold seminars on quality and statistical process control with engineers and leaders of industry within Japan. The lectures were a smashing success and demand for his services within Japan continued to grow throughout the 50s. Deming's teachings ended up being extremely influential in shaping the postwar Japanese economy and business culture and he became a national hero. Joseph Juran had a similar story a few years later, and he too taught the Japanese novel management techniques for quality management after receiving an invitation from JUSE. Deming and Juran's work was a crucial factor in Japan's post-WW2 economic miracle from 1950 to 1980, during which Japan rose from a disgraced and demolished nation to having the world's 2nd largest and most productive economy after the USA. Among the people enthusiastically following the advice were the leaders of Toyota.

Taiichi Ohno was the industrial engineer and later executive who was the primary architect of the Toyota Production System. While there is no evidence of a direct link between him and Deming, the TPS clearly came from the quality culture within Japan that Deming fostered almost single-handedly, and we can see that in the clear links between the principles Ohno inculcated into Toyota employees and managers. Additionally, Toyota has directly said that they were implementing Deming's teachings during their spectacular, disruptive rise to the top of the automotive industry. According to the Deming institute (link), "Years later, in 2005, Dr. Shoichiro Toyoda, Chairman and former President (1982-1999) of Toyota, accepted the American Society for Quality’s Deming Medal. In doing so, he offered: Every day I think about what he meant to us. Deming is the core of our management.” In that speech, Toyoda also said:
"…Dr. Deming came to Japan following World War II in order to teach industry leaders methods of statistical quality control, as well as to impart the significance of quality control in management and his overall management philosophy. He was an invaluable teacher…, playing an indispensable role in the development and revitalization of post-war Japan. Industrialists as well as academics earnestly began to study and implement Dr. Deming’s theories and philosophy. Dr. Deming soon became widely known not only as a brilliant theorist, but also as a kind and modest man. In 1951, the Deming Prize was founded in order to promote the widespread practice of quality control based on Dr. Deming’s philosophy. We at Toyota Motor Corporation introduced TQC in 1961, and in 1965 were awarded the Deming Application Prize…. As we continued to implement Dr. Deming’s teachings, we were able to both raise the level of quality of our products as well as enhance our operations on the corporate level. I believe that TMC today is a result of our continued efforts to implement positive change in pursuit of the Deming Prize..."
The revolutionary changes within Toyota in the 1950s transformed the company from a nearly bankrupt manufacturer of unreliable junky cars to the global champion of the automotive industry and arguably of the entire manufacturing sector of the global economy.

Notably, Ohno pushed Toyota to operate much like Tesla operates today, as illustrated by his "Ten Precepts":
  1. You are a cost. First reduce waste.
  2. First say, "I can do it." And try before everything.
  3. The workplace is a teacher. You can find answers only in the workplace.
  4. Do anything immediately. Starting something right now is the only way to win.
  5. Once you start something, persevere with it. Do not give up until you finish it.
  6. Explain difficult things in an easy-to-understand manner. Repeat things that are easy to understand.
  7. Waste is hidden. Do not hide it. Make problems visible.
  8. Valueless motions are equal to shortening one's life.
  9. Re-improve what was improved for further improvement.
  10. Wisdom is given equally to everybody. The point is whether one can exercise it.

This has clear parallels with how Tesla operates. For example:
  • Sleeping at the factory, putting all engineers next to the production line, requiring everyone to spend less time in meetings and more time directly engaging with the product
  • "I'd rather be optimistic and wrong than pessimistic and right."
  • "Optimism, pessimism, f*** that; we’re going to make it happen. As God is my bloody witness, I’m hell-bent on making it work."
  • "I don't ever give up. I mean, I'd have to be dead or completely incapacitated."
  • Jumping right into trying ideas immediately, almost pathologically impulsive innovation and impatience (e.g. Deciding to start the Boring Company, immediately planning for buying a boring machine and digging in the SpaceX parking lot, then actually starting work just five months later.)
  • Use of relatable analogies, humor, and repetitive mantras like "Prototypes are easy. Manufacturing is hard."
  • "Don’t delude yourself into thinking something’s working when it’s not, or you’re gonna get fixated on a bad solution."
  • "The worst mistake smart engineers make is optimizing something that shouldn't exist"
  • Continual optimization, making changes at astonishing rate
  • Empowering all employees to make suggestions, decisions and actually be heard by management

If Deming were still alive and could tour Giga Texas, I think he would be delighted beyond measure to see what Tesla has done and even more delighted to see it happening in America. After the war, Deming had been mostly ignored, shunned and scoffed at by leaders of American industry. 30 years passed after his first trip to Japan in 1950 until Americans finally were starting to panic about Japanese competition taking over. In 1980, NBC broadcast a highly influential documentary called "If Japan Can, Why Can't We?" that changed everything. If you only check out one video in this essay, this is the one I recommend the most. They talk specifically about Deming's role in part IV.


The credit Deming received from NBC for transforming Japan led to a surge in interest for his services in America and throughout the West and also led to his being hired at Ford, where a young tool-and-die maker named Sandy Munro was working. As it turned out, Sandy Munro ended up being directly mentored by Dr. Deming when both of them were employed at Ford in the 1980s. Few seem to understand the profound influence that Dr. Deming had on his protégé. In 1988, Deming directly recommended to Sandy that he leave behind Ford and its clueless, stubborn, arrogant managers and start his own consulting company, and thus Munro & Associates was born.


In 1982 Sandy Munro met Dr E. Deming and everything in his career changed. Dr. Deming’s ideas regarding reducing variation to increase quality resonated with Sandy and he became a zealot in reducing variation wherever he could find it. While working with Deming, Sandy started to think about how to tackle and eliminate variation and waste in the design phase that was unique to the product development process.
...
Sandy helped Ford utilize DFA [Design For Assembly] to save billions of dollars, improve quality and reduce development cycles during the early 1980’s, most notably in what became known as the “Taurus Shock” where Toyota who was starting to rise quickly and powerfully in the market due to the use of Kaizen and their TPS systems, found itself flabbergasted that in such a short time their competitors created such cost savings and innovation on the Ford Taurus design.
...
By a weird fluke, Sandy was moved into finance staff where he finally had the chance to look deeper into the heart of why poor design was such a large driver of poor manufacturing quality.

With great successes coming from the application of these tools in industry, Sandy was asked to speak alongside his teacher, adding to Deming’s speech that “As variation is reduced Quality will increase” (Deming), “all variation stems from design.” (Sandy Munro)
...
Truly, the underlying Deming spirit and philosophy is evident in all the Munro products and services.

Hear Sandy himself describe meeting Dr. Deming and what he learned when he became a self-proclaimed "Deming disciple":



So Munro and Associates was formed because of Deming's recommendation, with the explicit purpose of spreading Deming's message far and wide. Sandy basically copied his entire playbook from Deming’s principles, which are almost universally accepted as self-evident but which are, in practice, rarely executed with perfection. Here are some more Munro Live videos discussing Deming's influence on Sandy.



Elon & Sandy: Design Philosophy Parallels | PART 2

Deming's philosophy was foundational for building the Toyota Production System and Munro & Associates, and now the Munro team and the man himself are gushing with praise for Tesla's design for manufacturing and overall corporate management and culture.

Think about that.

These are obviously not just hollow opinions, because the Munro team have visited hundreds of factories and are tearing apart the cars and painstakingly reverse-engineering them, so they have all the experience and all the data needed to form an informed opinion. The fact that Sandy Munro is saying this is a primary reason why I'm confident in saying that Tesla is following the same timeless principles that other champions like Toyota have used, except Tesla has raised the bar for execution in a way nobody even thought was possible, making Tesla the new gold standard for quality control and lean manufacturing. Sandy (along with the late Joseph Juran) has been the American champion of Deming/Japanese-style lean design and quality engineering for the last three decades since Deming's death, and he is explicitly telling us that the way Elon does it at Tesla is the best example of quality engineering ever. This matters.

Some of Dr. Deming's final words on his deathbed in 1993 were, according to a USA Today article quoted by Wikipedia, "When asked, toward the end of his life, how he would wish to be remembered in the U.S., he replied, 'I probably won't even be remembered.' After a pause, he added, 'Well, maybe ... as someone who spent his life trying to keep America from committing suicide.'[41]" Sandy Munro, like his mentor, has been pleading for American industry to change and implement these principles for decades and has been mostly ignored or scoffed at until gaining celebrity and influence in his old age.

Let's examine more specifically why Tesla's principles align with Toyota's, and how Tesla is crushing Toyota and everybody else on execution.

Relentless optimization was championed by Taiichi Ohno and Toyota and is known today throughout industry by the Japanese word kaizen, which means going to the source of problems to see for oneself what is going wrong and always trying to implement improvements on a daily basis. It also involves grassroots improvement efforts coming from the people actually doing the work instead of command-and-control projects forced down from upper management onto the subordinates.

Kaizen is a daily process, the purpose of which goes beyond simple productivity improvement. It is also a process that, when done correctly, humanizes the workplace, eliminates overly hard work (muri), and teaches people how to perform experiments on their work using the scientific method and how to learn to spot and eliminate waste in business processes...People at all levels of an organization participate in kaizen, from the CEO down to janitorial staff, as well as external stakeholders when applicable...While kaizen (at Toyota) usually delivers small improvements, the culture of continual aligned small improvements and standardization yields large results in terms of overall improvement in productivity. This philosophy differs from the "command and control" improvement programs (e.g., Business Process Improvement) of the mid-20th century. Kaizen methodology includes making changes and monitoring results, then adjusting. Large-scale pre-planning and extensive project scheduling are replaced by smaller experiments, which can be rapidly adapted as new improvements are suggested.
Tesla definitely does kaizen, whether or not they label it as such.

At Tesla they leverage their software prowess to dramatically magnify the usefulness of kaizen. Tesla's two decades on infrastructure investment in automation and simulation tools have made what Joe Justice aptly calls dev ops for manufacturing. Design changes can be implemented on production vehicles quickly, in some cases almost immediately, because of simulation software has pre-validated that the change is likely to work.

Artificial intelligence using neural nets is fundamentally just fancy inferential statistics, because AI involves drawing inferences about a population based on a representative sample. Tesla's use of AI in the production As an outsider and a non-expert in AI, I'm limited in how much I can know about this. The difference is that neural nets can use brute force computation to run models that are vastly more sophisticated than what can be solved with classical statistics. The statistics of Deming and Shewhart still has a place for solving simpler problems, but for some more complicated challenges, we need machine learning. I speculate, as a company outsider, that Tesla is making extensive use of AI for end-item inspection and efficiently feeds that data back into machine learning models for. Joe Justice has claimed that Tesla does so as part of their digital self-management, but he's the only inside source I've heard it from.

They also benefit from having a leader who is autistic and trained in economics. The autism helps him not care much about saving face to maintain a high status position within a social hierarchy. The economics training helps him know to avoid the sunk cost fallacy and properly understand tradeoffs. Thus Elon is able to admit he is wrong easily, change direction when needed, and get everyone else at the company to follow his example.

One of Steven Mark Ryan's earliest YouTube videos captured this well in 2019.

"Elon Musk is a walking logic machine. Screw convention and what others think. Does it make sense? Is is right? Then do it! Not giving a s*** what others think and not living a life of convention can rub people the wrong way, hinder interpersonal relationships, and encourage unwarranted inputs from others, but none of this matters. What matters is doing what makes sense."


Tesla's kaizen efforts are also more productive due to compound effects of other aspects Tesla is getting right, including colocation of all factors of production into one gigafactory, vertical integration, and nearly zero inventory. More on this in later sections of this essay.

Tesla's goal is to minimize the amount of distance each atom travels from raw material extraction from ore to the customer receiving delivery. Even the selection of Austin for the new Tesla headquarters was partially because it's close to Boca Chica, so Elon doesn't have to travel as much.

Vertical integration allows Tesla to produce everything in one location, and a huge benefit of this colocation is that Tesla's people and material flow requires less transportation. For instance, Gigafactory Austin makes its own plastics, batteries, die castings, and seats all in one building. The pieces travel on the order of 1-100 meters to the subsequent production step. I mean, just look at the glorious efficiency in the factory layout shown below. Elon said while presenting this at the Cyber Rodeo Keynote:

You know a factory is advanced when it feels like an alien dreadnought landed. And so, the team's doing great work in Fremont, California, but we took a lot of lessons learned from that, where the buildings were all separate, there's a lot of movement between the buildings, and the thing that we thought made sense is to really think of it like a chip, like an integrated circuit. Combine everything into one package and now this is what you get. So this is a case of raw materials come in one side, they get formed into cell, they get formed into a pack, then we cast the front and rear body, the pack itself is structural, and out comes a finished product. So it's raw materials in one side, cars out the other side. (33:00 in presentation)

1662064468860.png


1662079610418.png


Giga Texas also has four floors. Most car factories have one floor. Multistory designs have been attempted before. For instance the Lingotto factory in Turin (link), Italy which made the first Fiat 500s. Ultimately the concept has not been successful, but when orchestrated well with robust software control, it can fundamentally improve production efficiency by reducing transportation waste. Building upward does increase construction cost, but it saves on land cost and allows more production capacity in a given site size. This speeds up pace of innovation due to proximity effects, and this is most important in the long run. Another simple advantage of going multistory is employees don't have to waste as much time walking between their parking space and their workstation. Less sprawl is better. The Boeing factory I worked at had this problem. Building jumbo jets inherently requires sprawl and going multistory would be total nonsense. The product is just huge and there's no way around it. But this unfortunately makes for a 7-15 minute walk for most people each way. That's a significant amount of time each day not used to its maximum potential.

Some companies ship their supplies halfway around the planet, or in some cases, as Elon has said is the case with some minerals like nickel, the material is shipped enough distance to circle the Earth multiple times over. Tesla will be getting lithium from places like Nevada and North Carolina and nickel from Brazil and Canada. Tesla will build some kind of battery or vehicle factory onsite in Indonesia using lithium from Australia and China. Etc.

The best part is no part. The best process is no process.
Reducing part count in a design typically results in reducing the total processing required to produce a manufactured object, because each part has associated process steps for ordering, fabrication, transportation, storage, installation, quality verification, and nonconformance rectification. Tesla has been drastically reducing non-value-added processing by reducing non-value-added part count and design complexity.

Examples:
  • Gigacastings replace hundreds of stamped parts and machining operations with two monolithic chunks of metal and a structural battery pack.
  • Munro & Associates teardowns have revealed a continual trend of simplification and part count reduction of Tesla vehicles, and as a result the amount of processing required is coming way down.
  • In response to chip shortage, Tesla has been rewriting firmware to reduce chip count and combine features into fewer microcontrollers.
  • Octovalve and heat pump cuts out unnecessary hoses, clamps, bottle, etc. for thermal management

The ideal stated by the TPS is to have a "pull" demand system with "just-in-time" logistics instead of a "push" supply system, meaning that there should be a cascading signal from the end customer upstream to the supply chain all the way down to Tier 3 suppliers, inspired by the way a grocery store restocks shelves in response to customer demand--but no more than that--then places orders for shipments of just what they need. The idea is that everything should be synchronized to customer demand. The opposite is a push production system that makes a certain amount of inventory and pushes it onto customers (internal or external), even if that means offering discounts to get rid of supply gluts or having shortages when demand gets high. The goal is also to avoid overproduction in individual process steps in the value stream, because if the next process downstream can't handle the output, then inventory piles up, cluttering the factory or causing an increase in scrap, interest and/or warehousing costs.

Tesla takes just-in-time to the extreme, but for one exception: that their customer order backlog is months or years long, forcing the customer to wait. This is temporary though, because Tesla can only scale so quickly, but eventually they will reach a scale where they can catch up on orders and deliver a completed car to a customer within days of them ordering it. Tesla's online, haggle-free, software-heavy approach to vehicle configuration, production and delivery drastically cuts the minimum lead time possible. This is how Tesla avoids inventory costs and write-downs on overstock and doesn't have to own a giant portfolio of real estate for traditional dealership parking lots.

Tesla has just 4 days of sales' worth of completed vehicle inventory on hand. No dealerships, customer order fulfilled within days of vehicle leaving factory.

In Q2 '22, Tesla had $8.1B in inventory on their balance. Tesla's 10-Q filing (link) for Q2 showed that the majority of the inventory was raw materials.

Note 6 – Inventor

Our inventory consisted of the following (in millions):


June 30,December 31,
20222021
Raw materials$4,949$2,816
Work in process1,3701,089
Finished goods (1)1,1851,277
Service parts604575
Total$8,108$5,757

(1)
Finished goods inventory includes vehicles in transit to fulfill customer orders, new vehicles available for sale, used vehicles, energy storage products and Solar Roof products available for sale.
We write-down inventory for any excess or obsolete inventories or when we believe that the net realizable value of inventories is less than the carrying value. During the three and six months ended June 30, 2022, we recorded write-downs of $23 million and $49 million, respectively, in Cost of revenues in the consolidated statements of operations. During the three and six months ended June 30, 2021, we recorded write-downs of $33 million and $56 million, respectively, in Cost of revenues in the consolidated statements of operations.
This is shockingly low, because their quarterly gross profit from operations was on average $4.8B in the first half of 2022. $8.1B / $4.8B * 3 months = 5 months of total cash flow from operations worth of inventory. This is very impressive when you consider that Tesla manufactures a lot of parts from scratch in-house and also owns their own distribution and service network for the customers, all of which means a significant amount of raw materials and work-in-process.

By comparison, Toyota's Q1 report showed they had a staggering $34B of inventory! However, this shot up recently due to supply chain shortages, so let's use their prior average inventory level of $23B. Toyota's quarterly revenue is usually ~$65B, so this is 23/65*91 = 32 days of sales' worth of total inventory for Toyota. This seems better than Tesla until you realize Toyota outsources a bunch of their raw material and work-in-process inventory cost to suppliers and doesn't have the dealership inventory.

I wrote about the benefits of Tesla's inventory minimization earlier this year.
Near-Zero Inventory
At Shanghai they literally do not even have a supply warehouse for materials and parts inventory, according to a few videos Tesla released in 2021 (links below).

Additionally, in the videos, I have noticed a conspicuous absence of almost any lineside inventory; hardly any shelving, carts, pallet queues, kanban bins or any other typical lineside storage can be seen. It seems they barely even use forklifts.

Semi trucks show up to the site, unload supplies on the side of the factory directly adjacent to the production line, and within probably an hour at most, the materials/components are on vehicles. They said at the time of the interview that they were processing nearly 2,000 shipping containers per day in the 97 loading docks. This works out to approximately 20 containers per bay each day, for an average cycle time of merely 70 minutes or so. Plus, they also localized the majority of their supply chain to nearby Chinese suppliers. In effect, this too reduces Tesla's inventory in their overall value chain because fewer parts and materials are in transit at any given moment.

Additionally, I'm astonished by the implications this all has on the level of quality control throughout the entire value chain that must have been achieved in order to enable having inventory buffers this low in the first place. Inventory exists fundamentally to accommodate variation by allowing the show to go on while an operation is deviating from an ideal state of continuous one-piece flow. If Tesla is producing thousands of cars per week from Shanghai with such low inventory, then this is strong evidence that they have drastically reduced variation and thus have increased first-pass quality to a level unheard of in the industry.
Massive Direct Cost Savings

Inventory reduction on this scale saves big money for Tesla.
  • Less capital required for work-in-progress
    • This increases overall return on capital, because the company earns nothing extra from this investment since inventory is non-value-added to the customer.
    • An ideal production line would, by magic, instantaneously transform raw materials to a completed widget in the customer's possession.
  • The upfront costs of the storage
    • Land
    • Building design and construction
    • Storage equipment: Bins, Drawers, Shelving, Racks, Pallets, Labels & Scanners, etc.
    • Transit equipment: Trucks, Forklifts, AGVs, etc.
  • Recurring warehouse operating and maintenance costs
    • Property tax
    • HVAC
    • Lighting
    • Building maintenance
    • Janitorial services
  • Extra transportation of materials
    • Moving items to and from storage locations is non-value added to the customer
    • Ongoing forklifts/truck/AGVs operating and maintenance expenses plus compensation for operators
    • More opportunity for a vehicle breakdown to disrupt production flow, introducing opportunity costs, overtime labor costs, expedited shipping costs, etc.
    • Safety costs (forklifts and trucks are one of the most common causes of severe industrial injuries, and even when no one is getting hurt, they still have to spend time and mental energy focusing on avoiding nearby vehicles)
  • Risk to the materials themselves
    • Damage in transit
    • Damage while stored (fire, leaking pipe or roof, earthquake, etc.)
    • Expiration (for materials like chemicals)
    • Theft
    • Accidental loss from misplacement

Design-Build Cycle Acceleration
Besides the fact that reducing inventory is obviously a fundamental goal of industrial engineering because it directly helps with lean just-in-time manufacturing, the advantage is compounded by the way Tesla uses extremely agile rapid-change design engineering practices. The shorter lag time from design modification to new parts arriving on the production line means pace of innovation is less often going to be constrained by this factor.

Reducing inventory also improves the areal and volumetric density of value-added steps within the building, which also improves the rate of innovation. Obviously, inventory takes up space on the factory floor. Having worked at the enormous Boeing campus in Everett, I saw how distance can be a big impediment to having tight feedback loops. For instance, the delivery flightline is a 20 minute walk from the factory or 5-10 minutes by tricycle. And most of the design engineers work in buildings that are 10-20 minutes of walking away from the factory and adjacent production facilities. Having the shortest possible distance between production steps facilitates communication, visual controls like Andons, and more productive gemba walks. In an interview from earlier this year, Tom Zhu, the president of Tesla China, said he spends at least an hour every single morning doing a full walk of the production line. The less distance he has to walk, the more information value he's getting out of each minute of walking. Also, psychology and time constraints dictate that in general, most people will not walk more than about 100 meters to solve a problem face to face rather than ignoring it or making a phone call. So, the density of people per 100 meter radius has a big impact on how fast ideas will spread, or be sparked by random conversations. Overall, simply having less junk in the way speeds up innovation.

Video References
These videos contain a lot more juicy manufacturing info than just inventory reduction stuff, but they clearly show the inventory system Tesla is using.

Inside Giga Shanghai new video w/ english subtitles

[English-Subtitled] Tesla Gigafactory Shanghai Exclusive Media Tour interview Part 1
and Part 2
and Part 3

Inside Gigafactory Shanghai - a Guided Tour of Tesla's most productive Factory

Selected Screenshots

dGHZDpc7md8zoRwM7TJ22IEoMmg0zrPKjXIhF6Y9MEjVmyoG945djNFOC9B9n5GsEgoawagJXoJ5sUeT6snPI8rmgSpJaV2vzhItfl2vfF_E-Td3EwGoffOt_Rs2IOZGCbR0_PM3

SJgkeBCgw2AwM0sR53zoRiBXaejlx9gjLiOXtqNbN5PM6aVVduAqipv1UbCKu0PEudaU6oucSFpouVXjGn7vIJOxhb58XJMH7qtW30jkD7yMXlmX0bnupqRWrfKHTJzbtQSVs0Pg





1645752479330.png


^ Supertub - Slide 42 of 2019 Impact Report


pWkYs6vhOnmJlwmycd0WRXaLKpoZJYd0vR3HEcWtHVOsrcvhjuO1ijk1MtKRQrf7GeaBfyRzB0ZJH6_fiW6f0BZRG7hSt0JO_mnVqnU3FFFJFLPDlGUkEHZkIirfboUyxntcabUz


^ Insertion and mating of completed dashboard assembly from feederline


oWE6pREv3Z0Pe-YNFRoZDbDLGNiDSsMIyZ4zcnA8MyHaeFh7l79Wpk2OrVEWynHP53b81Yh-gODs94Lhv9pObKELzrEKrEgQamYCU5vK9lZNHr4PEG1oH_9-rd94BNC1DdxHPWJy


^ Marriage of body-in-white and chassis/partially completed interior assembly, facilitated by elevators and AGVs

Tesla had a rough start with vehicle build quality but now leads the industry in the parts where it matters, and this is mainly due to radical simplification, design for manufacturability and ever-improving software and global team. As Elon has remarked, the tendency is for companies to slow down their rate of innovation as they scale, whereas incredibly Tesla's rate of innovation is increasing as they scale and it seems like they're just getting started.
    • Gigacastings have millimeter-level precision and consistently produce structures within tight geometrics tolerances because of lack of machining and lack of tolerance stackup of multiple small parts joined together. It's physically impossible to match this precision with traditional techniques.
    • According to an interview with a local Tesla China executive from Giga Shanghai, all of Tesla's wrenches are digital to ensure correct torque and recording data when it's applied.
    • Dry battery electrode deposition eliminates solvent and high heat in evaporation ovens, which should improve yield for that production step after the process is honed.
    • Tabless electrode design eliminates possibility of defects from traditional tab welding
    • Structural battery pack design allow for much easier and more ergonomic general assembly for the cabin.
    • Octovalve and supermanifold thermal management system architecture deletes a bunch of hose routing, clamping and termination operations
    • Munro teardowns have shown continual removal of screws and other threaded fasteners, increasing usage of plastic snap-fit parts, and in general lots of clever poka-yoke (mistake-proofing, designing part to be physically impossible to be installed incorrectly)
    • Tesla latest generation of motor and battery pack assembly almost entirely automated
There's not much to add to this section that I haven't already written about, because all of the wastes we want to reduce in lean design are connected. Deleted parts and processes can't be defective, because they don't exist. Quality is about reducing variation, and nothing can reduce variation more than eliminating a part or process. For the rest, the Tesla Automation team helps by making much of the production process be performed by reliable machines.

The greatest victory is that which requires no battle. -Sun Tzu

Back in the 1950s school of management thought, it was commonly accepted that labor relations conflicts were an unavoidable part of the manufacturing industry, that injuries were the unavoidable price to pay for production, and that manufacturing technicians were to do exactly as they were told and follow orders, just as their supervisors were expected to obey orders and so on up the "chain of command". People at the bottom were often berated and openly disrespected by management. Supervisors blamed employees for problems caused by variation, even though most the variation was caused by the system and the product design.

I think the greatest irony in this story is how the United States of the early Cold War years was obsessed with preserving free-market capitalism with generally free social policies, yet in that same era Americans tried to run their own private companies like miniature communist economies complete with centrally planned top-down control, suppression of opposing thought, enforcement of social conformity through at times arbitrary and capricious rules, and poorly aligned incentives that failed to encourage workers to give their best work, Big industrial companies also usually made the people working there feel like mindless drones whose ideas didn't matter and whose creative instinct was worthless, even though they were the people closest to the work with the best understanding of the processes. People didn't have authority to make decisions directly related to their work and there was a communication bottleneck caused by the chain of command edict that all communication to other managers should go through one's own manager. From a modern perspective we can easily see the obvious negative consequences that such poor management would have on safety, quality, productivity, employee retention rates, and most of all innovation, but in the 50s this was generally accepted as the normal way to run a company.

Toyota and the other Japanese industrial rising stars of the 1950s started looking at management differently. At Deming's recommendation, they started viewing their workforce as competent human beings who had ideas about how their work should look and what investments should be made to improve it. They believed if people were treated well and given room to improve their work, they will come up with ideas management never could have thought of. Managers offer ideas to, but it's much less about command and obedience than it was about collaboration. Controversially, Toyota even gave individual line employees the authority and responsibility to hit a button to stop the line whenever a quality issue arose, and then teams would swarm to solve issues directly on the production line. Taiichi Ohno would take chalk and draw a circle on the shop floor and instruct managers and engineers to stand in the circles for hours and observe production until they developed an understanding of what was going wrong for their team. Toyota also avoided having any layoffs for decades. On the technical side, they began capitalizing on Deming and Juran's methods for richer data collection and statistical analysis for root cause corrective action, prioritizing what to go after in kaizen efforts based on frequency of occurrence and magnitude of problem. This system inspired incredible employee loyalty and devotion, and it unleashed their creativity in the most efficient way possible because of the prioritization aspect. As a result, Toyota's pace of innovation, especially in manufacturing technology, dramatically outpaced everyone else in the car industry and they reaped the rewards of having more satisfying work that had fewer problems and earned big profits.

Elon spends the vast majority of his time these days right on the line with production teams, and observing for himself what was going on and how to help fix it. This is a very good sign. The Vice President of Tesla China also has said he spends a big portion of his day walking the line first thing in the morning after getting off the 6 AM call with the North American team. This is exactly what I want to see Tesla executives doing. There is no substitute for actually being next to the action. The amount of information you can absorb and process efficiently with your own senses is much greater than the information you can get from secondhand communication. The job of the plant management team is to set the team up for producing quality with systematic elimination of waste. So it's fundamentally better for the management team and other support teams like engineering to be at the production line where the poor quality and waste is happening and also where they can get rapid empirical feedback on whether their attempted innovations actually were working as hoped or not. This practice speeds up the maximum possible rate of innovation. In other words, it is waste of underutilized talent both in management and employees if employees are not authorized to make substantial decisions affecting their works, and also a case of underutilized talent in support groups wasting precious time and opportunity sitting in an office instead of being lineside. Since Tesla overwhelmingly avoids BS wastes of time like low-value email, useless meetings (most of them) and PowerPoint presentations, Tesla uses their talent to its maximum productive capacity.

Tesla empowers all the employees to make changes. Elon has mandated that the organization is flat and that everyone has an obligation to solve problems with their own brains and an obligation to speak to whomever is necessary for solving the problem. Also Tesla employees are required to experiment with changes in an attempt to innovate or else they will be fired, but on the flip side there's little to no punishment for sincerely trying to innovate and failing. This simple incentive structure produces a lot of extrinsic motivation for a ridiculous amount of innovation and unleashes the potential of the workforce, making them want to work long, hard hours because they're having fun and solving important problems at ludicrous speed, which generates excitement. The Tesla Anti-Handbook Handbook (link) emphasizes this point:

Trust
We give everyone who joins our team a lot of trust and responsibility. We operate with the assumption that everyone will do the right thing, including you...
...
Communication
Anyone at Tesla can and should email or talk to anyone else according to what they think is the fastest way to solve a problem for the benefit of the whole company. You can talk with your manager, you can talk to your manager's manager, you can talk directly to a VP in another department, you can talk to Elon--you can talk to anyone without anyone else's permission. Moreover, you should consider yourself obligated to do so until the right thing happens.

Job Duties
... Your #1 job -- everyone's #1 job -- is making this company a success. If you see opportunities to improve the way we do things, speak up even if these are outside your area of responsibility...make suggestions and share your ideas. Your good ideas mean nothing if you keep them to yourself.

Elon has also said everyone is supposed to think as though they were Chief Engineer. Workers don't waste time waiting for authorization to do things, and their ideas can be experimented with immediately instead of percolating through four layers of management in the "chain of command". They also minimize time wasted in meetings, even developing a new social norm of getting up and quietly leaving a meeting if you think you can add more value to the company by not being there. Joe Justice has harped on this point repeatedly.

Here's an excerpt from a big leaked email from 2018 sent by Elon to all employees (link):

Btw, here are a few productivity recommendations:

– Excessive meetings are the blight of big companies and almost always get worse over time. Please get of all large meetings, unless you’re certain they are providing value to the whole audience, in which case keep them very short.

– Also get rid of frequent meetings, unless you are dealing with an extremely urgent matter. Meeting frequency should drop rapidly once the urgent matter is resolved.

– Walk out of a meeting or drop off a call as soon as it is obvious you aren’t adding value. It is not rude to leave, it is rude to make someone stay and waste their time.

– Don’t use acronyms or nonsense words for objects, software or processes at Tesla. In general, anything that requires an explanation inhibits communication. We don’t want people to have to memorize a glossary just to function at Tesla.

– Communication should travel via the shortest path necessary to get the job done, not through the “chain of command”. Any manager who attempts to enforce chain of command communication will soon find themselves working elsewhere.

– A major source of issues is poor communication between depts. The way to solve this is allow free flow of information between all levels. If, in order to get something done between depts, an individual contributor has to talk to their manager, who talks to a director, who talks to a VP, who talks to another VP, who talks to a director, who talks to a manager, who talks to someone doing the actual work, then super dumb things will happen. It must be ok for people to talk directly and just make the right thing happen.

– In general, always pick common sense as your guide. If following a “company rule” is obviously ridiculous in a particular situation, such that it would make for a great Dilbert cartoon, then the rule should change.

If there is something you think should be done to make Tesla execute better or allow you to look forward to coming to work more (same thing in the long term), please send a note to [redacted]

Toyota is still the most profitable manufacturing company in history other than Samsung and TSMC, but Samsung and TSMC make their money selling high-margin semiconductor chips that require extremely advanced fabrication techniques and heavy capital investment in R&D and facilities. The semiconductor fabrication industry, unlike the car industry, has always tended towards having a couple dominant players making almost all the profits. TSMC and Samsung make gross margins of 40-50%. Toyota gets gross margin of more like 17-20%, which is exceptionally good for a manufacturer in a competitive industry selling their kind of volume.

Toyota has been an outlier in the global manufacturing industry for decades, but the days are numbered for Toyota's reign. Toyota epitomized the Japanese quality revolution and transformed how the world thinks about manufacturing and many other industries. Yet now, Tesla is mere months away from permanently surpassing Toyota in net income, despite Toyota having 5x greater unit volume, a 69-year head start, and much more old vehicles past their warranty expiration date generating high-margin sales for replacement parts. Tesla is beating Toyota by an enormous margin, and as they scale production in the next few years this gap will become ever more apparent.

Meanwhile, it appears that Toyota has calcified, lost their innovative edge, and are being mismanaged by their current CEO who probably was selected based on nepotism (considering that he is the grandson of Kiichiro Toyoda, the founder of Toyota's car division). As an outsider, I can only speculate. Perhaps complacency and haughtiness crept in over time. Akio Toyoda blamed it on excessive focus on growth in the 90s and early 00s (link). Overall, it seems to me that like so many individuals and organizations, Toyota has gotten away from disciplined application of the principles that made them successful in the first place, and they now find themselves in Stage 3, "Denial of Risk and Peril", of Jim Collins' organizational decay arc described in How the Mighty Fall.

1662055915010.png


What's even crazier is that Tesla is accomplishing this with only two products pulling almost all the weight, Model 3 and Model Y, and these are arguably just two variants of the same product, considering that they share 75% of their parts in common. The $65k+ Y is soon going to be outselling the Toyota Corolla, which has firmly held the top spot for half a century. Toyota has a diverse product portfolio serving a wide variety of market segments. Tesla still hasn't even begun to sell their best product, the Cybertruck, which is going to knock another half-century king off the throne: the Ford F-150. This all means that Tesla still has plenty of room to expand their range of offerings over time. Therefore, Tesla has a clear path to leveraging their design, manufacturing and distribution advantages across at least 10x as many car deliveries, because the S3XY lineup only addresses ~10% of the car and light truck market. Tesla's automotive division, excluding the FSD wildcard, is likely to be at least 10x the money-printing behemoth that Toyota has been.

The disruption that Toyota, in its heyday, brought to the automotive industry ended up radiating out to the entire global economy. What then will happen when Tesla makes an impact that dwarfs what Toyota did in the 20th century? Everybody eventually copied Toyota, or at least tried to, or failing that, at least paid lip service to trying. Toyota's success was too undeniable to ignore, and most people are followers anyway, including many of those who like to fancy themselves as leaders and independent thinkers. People, and especially kids, look to the high-status winners on top and attempt to emulate their behavior, often without even conscious awareness of the emulation. Tesla's success will be an order of magnitude more undeniable just on the car manufacturing business, not even counting the potential of FSD, Energy or Optimus. As the leader to whom everyone will be looking, Tesla will actually start influencing the entire macroeconomy, like Ford and Toyota before them. We can already see the beginnings of this with Ford and Volkswagen admitting that Tesla is kicking their asses and they need to copy Tesla as fast as possible. I expect this will eventually spread to all manufacturing companies.
 

Attachments

  • 1662015868487.png
    1662015868487.png
    2.3 MB · Views: 167
  • 1662064523518.png
    1662064523518.png
    1.4 MB · Views: 74
Last edited:
Forward Observing

If it’s Thursday evening, we were shopping at Costco (FYI ~ we own 500 +/- shares). We arrived al-a-cart (okay, cart with veggies). The guy that had just parked his Chevy SUV was excited to tell my wife he had just put his order in on a CyberTruck. My wife who has never been excited about our CyberTruck order boasted we hope to see ours by 2025.
All is good. Stock price should go up tomorrow.
 
I'm sure someone has done the math, but one of the better odd decisions Tesla made was to take all leases back for model 3+Y and not allow the customer to buy them out. The sheer margin they are making after customers paying the cost of the car, then giving it back to Tesla to sell again for 10K more than the first time is amazing. Not to mention they are pretty much forcing FSD purchases on anything used. Tesla doesnt lease 3+Y to the same degree as other luxury brands...but its nearly pure profit for them now.
 
I’m surprised there hasn’t been more discussion on the China August numbers (77k). Giga Shanghai is at almost at 1 million Vehicles per year rate now and will easily exceed it by the end of the year with the recent upgrades. Berlin and Austin are less than two years behind. That puts us around 4 million per year by the end of 2024.
DE3A2E5A-9B7B-47D2-AE4D-89EF6A084AAF.jpeg
 
Uber gets it. 10k more EVs in London…..vehicle of choice: Model 3.

I chat to the Uber drivers here in London quite a bit (I take 1-2 trips a week) and many of them have already gone electric. The feedback I have received from them is that they save £1k-£2k a month switching from ICE to EV due to lower fuel costs, lower rego/taxes, no ULEZ fees and lower maintenance. They also get a slightly higher price per mile for "Uber Green" trips. That's £12k-£24k a year extra in their pocket - which is an amazing benefit given the salary of the average Uber driver. Electricity costs are skyrocketing here at the moment so the above numbers may have changed a little.

They do mention some hassles around charging from time to time but that most days they can get their work done on a single trip.

The most popular EVs appear to be Konas and Niros, although I have started seeing more Chinese EVs lately.
 
Tesla, Toyota, and The Next Generation of Lean Manufacturing
The Toyota Production System and the Toyota Way revolutionized not only the automotive industry but also the entire global manufacturing sector. I've worked as a quality engineer at Boeing and another major industrial company in a different industry. Like many companies, my employers designed their own production systems basically as rebranded variants of the TPS adapted for their own purposes. In manufacturing, it is common to use Japanese words taken straight from the TPS, like andon, poka-yoke, kanban, gemba, kaizen, muri, and muda. The propulsion shop at my Boeing plant even had posters on the wall of Taiichi Ohno and some of his most famous sayings. The prosperity we live in today is in no small part due to the contributions of the Toyota Production System and the Toyota Way to changing the way businesses are run.

The fact that Tesla is now blowing the TPS out of the water is, therefore, astonishing. However, most of what Tesla is doing in their manufacturing efforts is not really new fundamentally (and the same goes for SpaceX and Boring). Tesla's production system is based on the same principles established in the TPS and Toyota Way and raising the standard of execution on them in practice to new levels never seen before, while massively amplifying the effect with modern computer technology. I don't know if Tesla is intentionally copying Toyota and pulverizing them at their own game, but in effect that is what's happening. There has been no sign of Tesla using Toyota's terminology and they've never said they're learning from Toyota, so it's quite possible Elon has reinvented the wheel on a lot of this by failing a bunch and learning quickly from his mistakes, but in the end it doesn't really matter why he follows the principles because either way the principles work and the outcome is the same.


Elon came up with some catchier phrasing for the principles but the foundations of Lean Manufacturing have been known for longer than he's been alive. For instance, compare Toyota's "8 Wastes" to principles Tesla uses for industrial engineering.
  1. Waste of overproduction (largest waste)
  2. Waste of time on hand (waiting)
  3. Waste of transportation
  4. Waste of processing itself
  5. Waste of excess inventory
  6. Waste of movement
  7. Waste of making defective products
  8. Waste of underutilized workers
Reducing these forms of waste is the first principle of industrial engineering. This is self-evident. If you expend valuable resources doing any these 8 things that the customer doesn't value (or values negatively), then by definition you are wasting resources. An effective organization transforms inputs of a certain value into outputs of a higher value.

The key problem with adopting Lean and attempting to reduce waste is not in understanding the framework, but rather the problem is, and always has been, psychological factors, especially within management and even more so in top management. Ego, fear, lack of teamwork, arrogance, laziness, greed, distrust, resistance to change, short-termism, and other age-old human flaws stand in the way of the right stuff happening. I believe that this is fundamentally where Tesla and Elon Musk excel, and why they are now stomping on Toyota. Manufacturing will be Tesla's long-term competitive advantage and in this essay I want to summarize my thesis for why.

After WW2, American armed forces and other Allied powers continued occupying Japan, helping rebuild Japanese society and industry, in part due to a desire to increase American influence in East Asia and prevent Japan from descending into communism under Soviet shadow control, as eventually happened in North Korea, China, Mongolia, Vietnam and many other Southeast Asian nations. In 1947, General Douglas MacArthur requested that a statistician named Dr. William Edwards Deming be transferred to Japan to help the government census in Japan and to help with rebuilding Japan in general. Deming had an academic background in physics and mathematics at Yale and prior to his first trip to Japan he had been working for the US Census Bureau helping them implement much better internal quality processes and statistical sampling techniques.

Twenty years prior, Deming had learned radical new techniques and management frameworks while working at Bell Labs with Walter Shewhart, a fellow physicist who had pioneered modern statistical process control. As physicists who came of age in an era during which the very philosophical foundations of science were being questioned question as theories like Relativity and Quantum Mechanics were coming into vogue, both men had a keen intellectual interest in epistemology, the study of where knowledge comes from. As scientists by training, both also possessed the skills and mathematical understand to actually apply the Scientific Method correctly. They transferred this knowledge and mindset into their ideas for improving organizational processes with rigorous empirical techniques. Deming was enamored with Shewhart's ideas and helped Shewhart develop them further. By the time the war happened, Deming had gained a substantial reputation, especially within the federal government after having improved efficiency at several departments and especially during the war when he trained many American workers on quality improvement.

At the time, Japan had a well-deserved reputation for producing junk. They had serious problems with quality and low productivity, and so the Japanese Union of Scientists and Engineers had keen interest in improving. Meanwhile, North America was the only area of the world with advanced industrial capacity that hadn't been damaged by the war, and America in particular had undergone a revolution in quality management and statistical process control in the years leading up to WWII and then especially during the war when production was mission-critical and they had to figure out how to quickly get women trained and up to speed on the production lines. Many of the techniques of acceptance sampling and statistical process control developed during this period are still in use today, and I personally used them while at Boeing. Deming himself was quite influential in the spread of these techniques for supporting the war effort. The Japanese, humbled after losing the war, were ready to learn from what others, particularly the Americans, were doing so much better than they were. Meanwhile, the Americans took a turn away from what had helped make the war production effort so successful, and they began focusing less on quality and more on piece cost reduction and cranking out maximum production quantity at the expense of everything else. This appeared to work for a while, largely because of the USA having overwhelmingly the best postwar industrial capacity which fed the immense demand from American allies whose factories and roads had been leveled by bombs, and also because of the cheap domestic American oil and coal of the 20th century. Japanese engineers and industrial leaders frequently visited America during this time and absorbed knowledge from industrial titans like Ford, Bell, and Westinghouse, and even American retail and grocery stores.

So, while working for the US government rebuilding efforts in Japan, in 1950 Deming received requests from the Japanese Union of Scientists and Engineers to give lectures and hold seminars on quality and statistical process control with engineers and leaders of industry within Japan. The lectures were a smashing success and demand for his services within Japan continued to grow throughout the 50s. Deming's teachings ended up being extremely influential in shaping the postwar Japanese economy and business culture and he became a national hero. Joseph Juran had a similar story a few years later, and he too taught the Japanese novel management techniques for quality management after receiving an invitation from JUSE. Deming and Juran's work was a crucial factor in Japan's post-WW2 economic miracle from 1950 to 1980, during which Japan rose from a disgraced and demolished nation to having the world's 2nd largest and most productive economy after the USA. Among the people enthusiastically following the advice were the leaders of Toyota.

Taiichi Ohno was the industrial engineer and later executive who was the primary architect of the Toyota Production System. While there is no evidence of a direct link between him and Deming, the TPS clearly came from the quality culture within Japan that Deming fostered almost single-handedly, and we can see that in the clear links between the principles Ohno inculcated into Toyota employees and managers. Additionally, Toyota has directly said that they were implementing Deming's teachings during their spectacular, disruptive rise to the top of the automotive industry. According to the Deming institute (link), "Years later, in 2005, Dr. Shoichiro Toyoda, Chairman and former President (1982-1999) of Toyota, accepted the American Society for Quality’s Deming Medal. In doing so, he offered: Every day I think about what he meant to us. Deming is the core of our management.” In that speech, Toyoda also said:

The revolutionary changes within Toyota in the 1950s transformed the company from a nearly bankrupt manufacturer of unreliable junky cars to the global champion of the automotive industry and arguably of the entire manufacturing sector of the global economy.

Notably, Ohno pushed Toyota to operate much like Tesla operates today, as illustrated by his "Ten Precepts":


This has clear parallels with how Tesla operates. For example:
  • Sleeping at the factory, putting all engineers next to the production line, requiring everyone to spend less time in meetings and more time directly engaging with the product
  • "I'd rather be optimistic and wrong than pessimistic and right."
  • "Optimism, pessimism, f*** that; we’re going to make it happen. As God is my bloody witness, I’m hell-bent on making it work."
  • "I don't ever give up. I mean, I'd have to be dead or completely incapacitated."
  • Jumping right into trying ideas immediately, almost pathologically impulsive innovation and impatience (e.g. Deciding to start the Boring Company, immediately planning for buying a boring machine and digging in the SpaceX parking lot, then actually starting work just five months later.)
  • Use of relatable analogies, humor, and repetitive mantras like "Prototypes are easy. Manufacturing is hard."
  • "Don’t delude yourself into thinking something’s working when it’s not, or you’re gonna get fixated on a bad solution."
  • "The worst mistake smart engineers make is optimizing something that shouldn't exist"
  • Continual optimization, making changes at astonishing rate
  • Empowering all employees to make suggestions, decisions and actually be heard by management

If Deming were still alive and could tour Giga Texas, I think he would be delighted beyond measure to see what Tesla has done and even more delighted to see it happening in America. After the war, Deming had been mostly ignored, shunned and scoffed at by leaders of American industry. 30 years passed after his first trip to Japan in 1950 until Americans finally were starting to panic about Japanese competition taking over. In 1980, NBC broadcast a highly influential documentary called "If Japan Can, Why Can't We?" that changed everything. If you only check out one video in this essay, this is the one I recommend the most. They talk specifically about Deming's role in part IV.


The credit Deming received from NBC for transforming Japan led to a surge in interest for his services in America and throughout the West and also led to his being hired at Ford, where a young tool-and-die maker named Sandy Munro was working. As it turned out, Sandy Munro ended up being directly mentored by Dr. Deming when both of them were employed at Ford in the 1980s. Few seem to understand the profound influence that Dr. Deming had on his protégé. In 1988, Deming directly recommended to Sandy that he leave behind Ford and its clueless, stubborn, arrogant managers and start his own consulting company, and thus Munro & Associates was born.




Hear Sandy himself describe meeting Dr. Deming and what he learned when he became a self-proclaimed "Deming disciple":



So Munro and Associates was formed because of Deming's recommendation, with the explicit purpose of spreading Deming's message far and wide. Sandy basically copied his entire playbook from Deming’s principles, which are almost universally accepted as self-evident but which are, in practice, rarely executed with perfection. Here are some more Munro Live videos discussing Deming's influence on Sandy.



Elon & Sandy: Design Philosophy Parallels | PART 2

Deming's philosophy was foundational for building the Toyota Production System and Munro & Associates, and now the Munro team and the man himself are gushing with praise for Tesla's design for manufacturing and overall corporate management and culture.

Think about that.

These are obviously not just hollow opinions, because the Munro team have visited hundreds of factories and are tearing apart the cars and painstakingly reverse-engineering them, so they have all the experience and all the data needed to form an informed opinion. The fact that Sandy Munro is saying this is a primary reason why I'm confident in saying that Tesla is following the same timeless principles that other champions like Toyota have used, except Tesla has raised the bar for execution in a way nobody even thought was possible, making Tesla the new gold standard for quality control and lean manufacturing. Sandy (along with the late Joseph Juran) has been the American champion of Deming/Japanese-style lean design and quality engineering for the last three decades since Deming's death, and he is explicitly telling us that the way Elon does it at Tesla is the best example of quality engineering ever. This matters.

Some of Dr. Deming's final words on his deathbed in 1993 were, according to a USA Today article quoted by Wikipedia, "When asked, toward the end of his life, how he would wish to be remembered in the U.S., he replied, 'I probably won't even be remembered.' After a pause, he added, 'Well, maybe ... as someone who spent his life trying to keep America from committing suicide.'[41]" Sandy Munro, like his mentor, has been pleading for American industry to change and implement these principles for decades and has been mostly ignored or scoffed at until gaining celebrity and influence in his old age.

Let's examine more specifically why Tesla's principles align with Toyota's, and how Tesla is crushing Toyota and everybody else on execution.

Relentless optimization was championed by Taiichi Ohno and Toyota and is known today throughout industry by the Japanese word kaizen, which means going to the source of problems to see for oneself what is going wrong and always trying to implement improvements on a daily basis. It also involves grassroots improvement efforts coming from the people actually doing the work instead of command-and-control projects forced down from upper management onto the subordinates.


Tesla definitely does kaizen, whether or not they label it as such.

At Tesla they leverage their software prowess to dramatically magnify the usefulness of kaizen. Tesla's two decades on infrastructure investment in automation and simulation tools have made what Joe Justice aptly calls dev ops for manufacturing. Design changes can be implemented on production vehicles quickly, in some cases almost immediately, because of simulation software has pre-validated that the change is likely to work.

Artificial intelligence using neural nets is fundamentally just fancy inferential statistics, because AI involves drawing inferences about a population based on a representative sample. Tesla's use of AI in the production As an outsider and a non-expert in AI, I'm limited in how much I can know about this. The difference is that neural nets can use brute force computation to run models that are vastly more sophisticated than what can be solved with classical statistics. The statistics of Deming and Shewhart still has a place for solving simpler problems, but for some more complicated challenges, we need machine learning. I speculate, as a company outsider, that Tesla is making extensive use of AI for end-item inspection and efficiently feeds that data back into machine learning models for. Joe Justice has claimed that Tesla does so as part of their digital self-management, but he's the only inside source I've heard it from.

They also benefit from having a leader who is autistic and trained in economics. The autism helps him not care much about saving face to maintain a high status position within a social hierarchy. The economics training helps him know to avoid the sunk cost fallacy and properly understand tradeoffs. Thus Elon is able to admit he is wrong easily, change direction when needed, and get everyone else at the company to follow his example.

One of Steven Mark Ryan's earliest YouTube videos captured this well in 2019.




Tesla's kaizen efforts are also more productive due to compound effects of other aspects Tesla is getting right, including colocation of all factors of production into one gigafactory, vertical integration, and nearly zero inventory. More on this in later sections of this essay.

Tesla's goal is to minimize the amount of distance each atom travels from raw material extraction from ore to the customer receiving delivery. Even the selection of Austin for the new Tesla headquarters was partially because it's close to Boca Chica, so Elon doesn't have to travel as much.

Vertical integration allows Tesla to produce everything in one location, and a huge benefit of this colocation is that Tesla's people and material flow requires less transportation. For instance, Gigafactory Austin makes its own plastics, batteries, die castings, and seats all in one building. The pieces travel on the order of 1-100 meters to the subsequent production step. I mean, just look at the glorious efficiency in the factory layout shown below. Elon said while presenting this at the Cyber Rodeo Keynote:



View attachment 847965

View attachment 848050

Giga Texas also has four floors. Most car factories have one floor. Multistory designs have been attempted before. For instance the Lingotto factory in Turin (link), Italy which made the first Fiat 500s. Ultimately the concept has not been successful, but when orchestrated well with robust software control, it can fundamentally improve production efficiency by reducing transportation waste. Building upward does increase construction cost, but it saves on land cost and allows more production capacity in a given site size. This speeds up pace of innovation due to proximity effects, and this is most important in the long run. Another simple advantage of going multistory is employees don't have to waste as much time walking between their parking space and their workstation. Less sprawl is better. The Boeing factory I worked at had this problem. Building jumbo jets inherently requires sprawl and going multistory would be total nonsense. The product is just huge and there's no way around it. But this unfortunately makes for a 7-15 minute walk for most people each way. That's a significant amount of time each day not used to its maximum potential.

Some companies ship their supplies halfway around the planet, or in some cases, as Elon has said is the case with some minerals like nickel, the material is shipped enough distance to circle the Earth multiple times over. Tesla will be getting lithium from places like Nevada and North Carolina and nickel from Brazil and Canada. Tesla will build some kind of battery or vehicle factory onsite in Indonesia using lithium from Australia and China. Etc.

Reducing part count in a design typically results in reducing the total processing required to produce a manufactured object, because each part has associated process steps for ordering, fabrication, transportation, storage, installation, quality verification, and nonconformance rectification. Tesla has been drastically reducing non-value-added processing by reducing non-value-added part count and design complexity.

Examples:
  • Gigacastings replace hundreds of stamped parts and machining operations with two monolithic chunks of metal and a structural battery pack.
  • Munro & Associates teardowns have revealed a continual trend of simplification and part count reduction of Tesla vehicles, and as a result the amount of processing required is coming way down.
  • In response to chip shortage, Tesla has been rewriting firmware to reduce chip count and combine features into fewer microcontrollers.
  • Octovalve and heat pump cuts out unnecessary hoses, clamps, bottle, etc. for thermal management

The ideal stated by the TPS is to have a "pull" demand system with "just-in-time" logistics instead of a "push" supply system, meaning that there should be a cascading signal from the end customer upstream to the supply chain all the way down to Tier 3 suppliers, inspired by the way a grocery store restocks shelves in response to customer demand--but no more than that--then places orders for shipments of just what they need. The idea is that everything should be synchronized to customer demand. The opposite is a push production system that makes a certain amount of inventory and pushes it onto customers (internal or external), even if that means offering discounts to get rid of supply gluts or having shortages when demand gets high. The goal is also to avoid overproduction in individual process steps in the value stream, because if the next process downstream can't handle the output, then inventory piles up, cluttering the factory or causing an increase in scrap, interest and/or warehousing costs.

Tesla takes just-in-time to the extreme, but for one exception: that their customer order backlog is months or years long, forcing the customer to wait. This is temporary though, because Tesla can only scale so quickly, but eventually they will reach a scale where they can catch up on orders and deliver a completed car to a customer within days of them ordering it. Tesla's online, haggle-free, software-heavy approach to vehicle configuration, production and delivery drastically cuts the minimum lead time possible. This is how Tesla avoids inventory costs and write-downs on overstock and doesn't have to own a giant portfolio of real estate for traditional dealership parking lots.

Tesla has just 4 days of sales' worth of completed vehicle inventory on hand. No dealerships, customer order fulfilled within days of vehicle leaving factory.

In Q2 '22, Tesla had $8.1B in inventory on their balance. Tesla's 10-Q filing (link) for Q2 showed that the majority of the inventory was raw materials.


This is shockingly low, because their quarterly gross profit from operations was on average $4.8B in the first half of 2022. $8.1B / $4.8B * 3 months = 5 months of total cash flow from operations worth of inventory. This is very impressive when you consider that Tesla manufactures a lot of parts from scratch in-house and also owns their own distribution and service network for the customers, all of which means a significant amount of raw materials and work-in-process.

By comparison, Toyota's Q1 report showed they had a staggering $34B of inventory! However, this shot up recently due to supply chain shortages, so let's use their prior average inventory level of $23B. Toyota's quarterly revenue is usually ~$65B, so this is 23/65*91 = 32 days of sales' worth of total inventory for Toyota. This seems better than Tesla until you realize Toyota outsources a bunch of their raw material and work-in-process inventory cost to suppliers and doesn't have the dealership inventory.

I wrote about the benefits of Tesla's inventory minimization earlier this year.

Tesla had a rough start with vehicle build quality but now leads the industry in the parts where it matters, and this is mainly due to radical simplification, design for manufacturability and ever-improving software and global team. As Elon has remarked, the tendency is for companies to slow down their rate of innovation as they scale, whereas incredibly Tesla's rate of innovation is increasing as they scale and it seems like they're just getting started.
    • Gigacastings have millimeter-level precision and consistently produce structures within tight geometrics tolerances because of lack of machining and lack of tolerance stackup of multiple small parts joined together. It's physically impossible to match this precision with traditional techniques.
    • According to an interview with a local Tesla China executive from Giga Shanghai, all of Tesla's wrenches are digital to ensure correct torque and recording data when it's applied.
    • Dry battery electrode deposition eliminates solvent and high heat in evaporation ovens, which should improve yield for that production step after the process is honed.
    • Tabless electrode design eliminates possibility of defects from traditional tab welding
    • Structural battery pack design allow for much easier and more ergonomic general assembly for the cabin.
    • Octovalve and supermanifold thermal management system architecture deletes a bunch of hose routing, clamping and termination operations
    • Munro teardowns have shown continual removal of screws and other threaded fasteners, increasing usage of plastic snap-fit parts, and in general lots of clever poka-yoke (mistake-proofing, designing part to be physically impossible to be installed incorrectly)
    • Tesla latest generation of motor and battery pack assembly almost entirely automated
There's not much to add to this section that I haven't already written about, because all of the wastes we want to reduce in lean design are connected. Deleted parts and processes can't be defective, because they don't exist. Quality is about reducing variation, and nothing can reduce variation more than eliminating a part or process. For the rest, the Tesla Automation team helps by making much of the production process be performed by reliable machines.

Back in the 1950s school of management thought, it was commonly accepted that labor relations conflicts were an unavoidable part of the manufacturing industry, that injuries were the unavoidable price to pay for production, and that manufacturing technicians were to do exactly as they were told and follow orders, just as their supervisors were expected to obey orders and so on up the "chain of command". People at the bottom were often berated and openly disrespected by management. Supervisors blamed employees for problems caused by variation, even though most the variation was caused by the system and the product design.

I think the greatest irony in this story is how the United States of the early Cold War years was obsessed with preserving free-market capitalism with generally free social policies, yet in that same era Americans tried to run their own private companies like miniature communist economies complete with centrally planned top-down control, suppression of opposing thought, enforcement of social conformity through at times arbitrary and capricious rules, and poorly aligned incentives that failed to encourage workers to give their best work, Big industrial companies also usually made the people working there feel like mindless drones whose ideas didn't matter and whose creative instinct was worthless, even though they were the people closest to the work with the best understanding of the processes. People didn't have authority to make decisions directly related to their work and there was a communication bottleneck caused by the chain of command edict that all communication to other managers should go through one's own manager. From a modern perspective we can easily see the obvious negative consequences that such poor management would have on safety, quality, productivity, employee retention rates, and most of all innovation, but in the 50s this was generally accepted as the normal way to run a company.

Toyota and the other Japanese industrial rising stars of the 1950s started looking at management differently. At Deming's recommendation, they started viewing their workforce as competent human beings who had ideas about how their work should look and what investments should be made to improve it. They believed if people were treated well and given room to improve their work, they will come up with ideas management never could have thought of. Managers offer ideas to, but it's much less about command and obedience than it was about collaboration. Controversially, Toyota even gave individual line employees the authority and responsibility to hit a button to stop the line whenever a quality issue arose, and then teams would swarm to solve issues directly on the production line. Taiichi Ohno would take chalk and draw a circle on the shop floor and instruct managers and engineers to stand in the circles for hours and observe production until they developed an understanding of what was going wrong for their team. Toyota also avoided having any layoffs for decades. On the technical side, they began capitalizing on Deming and Juran's methods for richer data collection and statistical analysis for root cause corrective action, prioritizing what to go after in kaizen efforts based on frequency of occurrence and magnitude of problem. This system inspired incredible employee loyalty and devotion, and it unleashed their creativity in the most efficient way possible because of the prioritization aspect. As a result, Toyota's pace of innovation, especially in manufacturing technology, dramatically outpaced everyone else in the car industry and they reaped the rewards of having more satisfying work that had fewer problems and earned big profits.

Elon spends the vast majority of his time these days right on the line with production teams, and observing for himself what was going on and how to help fix it. This is a very good sign. The Vice President of Tesla China also has said he spends a big portion of his day walking the line first thing in the morning after getting off the 6 AM call with the North American team. This is exactly what I want to see Tesla executives doing. There is no substitute for actually being next to the action. The amount of information you can absorb and process efficiently with your own senses is much greater than the information you can get from secondhand communication. The job of the plant management team is to set the team up for producing quality with systematic elimination of waste. So it's fundamentally better for the management team and other support teams like engineering to be at the production line where the poor quality and waste is happening and also where they can get rapid empirical feedback on whether their attempted innovations actually were working as hoped or not. This practice speeds up the maximum possible rate of innovation. In other words, it is waste of underutilized talent both in management and employees if employees are not authorized to make substantial decisions affecting their works, and also a case of underutilized talent in support groups wasting precious time and opportunity sitting in an office instead of being lineside. Since Tesla overwhelmingly avoids BS wastes of time like low-value email, useless meetings (most of them) and PowerPoint presentations, Tesla uses their talent to its maximum productive capacity.

Tesla empowers all the employees to make changes. Elon has mandated that the organization is flat and that everyone has an obligation to solve problems with their own brains and an obligation to speak to whomever is necessary for solving the problem. Also Tesla employees are required to experiment with changes in an attempt to innovate or else they will be fired, but on the flip side there's little to no punishment for sincerely trying to innovate and failing. This simple incentive structure produces a lot of extrinsic motivation for a ridiculous amount of innovation and unleashes the potential of the workforce, making them want to work long, hard hours because they're having fun and solving important problems at ludicrous speed, which generates excitement. The Tesla Anti-Handbook Handbook (link) emphasizes this point:



Elon has also said everyone is supposed to think as though they were Chief Engineer. Workers don't waste time waiting for authorization to do things, and their ideas can be experimented with immediately instead of percolating through four layers of management in the "chain of command". They also minimize time wasted in meetings, even developing a new social norm of getting up and quietly leaving a meeting if you think you can add more value to the company by not being there. Joe Justice has harped on this point repeatedly.

Here's an excerpt from a big leaked email from 2018 sent by Elon to all employees (link):

Toyota is still the most profitable manufacturing company in history other than Samsung and TSMC, but Samsung and TSMC make their money selling high-margin semiconductor chips that require extremely advanced fabrication techniques and heavy capital investment in R&D and facilities. The semiconductor fabrication industry, unlike the car industry, has always tended towards having a couple dominant players making almost all the profits. TSMC and Samsung make gross margins of 40-50%. Toyota gets gross margin of more like 17-20%, which is exceptionally good for a manufacturer in a competitive industry selling their kind of volume.

Toyota has been an outlier in the global manufacturing industry for decades, but the days are numbered for Toyota's reign. Toyota epitomized the Japanese quality revolution and transformed how the world thinks about manufacturing and many other industries. Yet now, Tesla is mere months away from permanently surpassing Toyota in net income, despite Toyota having 5x greater unit volume, a 69-year head start, and much more old vehicles past their warranty expiration date generating high-margin sales for replacement parts. Tesla is beating Toyota by an enormous margin, and as they scale production in the next few years this gap will become ever more apparent.

Meanwhile, it appears that Toyota has calcified, lost their innovative edge, and are being mismanaged by their current CEO who probably was selected based on nepotism (considering that he is the grandson of Kiichiro Toyoda, the founder of Toyota's car division). As an outsider, I can only speculate. Perhaps complacency and haughtiness crept in over time. Akio Toyoda blamed it on excessive focus on growth in the 90s and early 00s (link). Overall, it seems to me that like so many individuals and organizations, Toyota has gotten away from disciplined application of the principles that made them successful in the first place, and they now find themselves in Stage 3, "Denial of Risk and Peril", of Jim Collins' organizational decay arc described in How the Mighty Fall.

View attachment 847926

What's even crazier is that Tesla is accomplishing this with only two products pulling almost all the weight, Model 3 and Model Y, and these are arguably just two variants of the same product, considering that they share 75% of their parts in common. The $65k+ Y is soon going to be outselling the Toyota Corolla, which has firmly held the top spot for half a century. Toyota has a diverse product portfolio serving a wide variety of market segments. Tesla still hasn't even begun to sell their best product, the Cybertruck, which is going to knock another half-century king off the throne: the Ford F-150. This all means that Tesla still has plenty of room to expand their range of offerings over time. Therefore, Tesla has a clear path to leveraging their design, manufacturing and distribution advantages across at least 10x as many car deliveries, because the S3XY lineup only addresses ~10% of the car and light truck market. Tesla's automotive division, excluding the FSD wildcard, is likely to be at least 10x the money-printing behemoth that Toyota has been.

The disruption that Toyota, in its heyday, brought to the automotive industry ended up radiating out to the entire global economy. What then will happen when Tesla makes an impact that dwarfs what Toyota did in the 20th century? Everybody eventually copied Toyota, or at least tried to, or failing that, at least paid lip service to trying. Toyota's success was too undeniable to ignore, and most people are followers anyway, including many of those who like to fancy themselves as leaders and independent thinkers. People, and especially kids, look to the high-status winners on top and attempt to emulate their behavior, often without even conscious awareness of the emulation. Tesla's success will be an order of magnitude more undeniable just on the car manufacturing business, not even counting the potential of FSD, Energy or Optimus. As the leader to whom everyone will be looking, Tesla will actually start influencing the entire macroeconomy, like Ford and Toyota before them. We can already see the beginnings of this with Ford and Volkswagen admitting that Tesla is kicking their asses and they need to copy Tesla as fast as possible. I expect this will eventually spread to all manufacturing companies.

What an excellent post! That must have taken a lot of work to put together.

There are a couple of things, which are touched on, but I think, not given enough weight.

Tesla own the entire product life, from production through to recycling. This allows optimisations for the buying experience, servicing, updates (mainly software but also e.g. HW3), insurance, second hand market, and end of life recycling. It not only allows optimisations, but gives Tesla a strong incentive to make them. We see this in Tesla's strong emphasis on reliability and product longevity.

By high vertical integration it is much easier to combine sub-assemblies, or move functionality from one sub-assembly to another. e.g. octovalue, e.g. combining functions of microcontrollers. A supplier for widget "A" can optimise and continuously improve within the limits of the boundaries of the interfaces to that widget, but it is almost impossible for them to co-optimise with widget "B" made by a different supplier at a different factory. This is especially true for parts that come from multiple suppliers (so that there is no dependency on a supplier that may fail). Interfaces between teams and especially companies add impedence to communication and provide perverse incentives (if we optimise this part, or make it cheaper or move its functionality elsewhere, or eliminate its functionality, then our team will be less important, the suppliers revenue will be less, all our work on the eliminated part will be thrown away).

By moving functionality to software there can be continuing product improvement even after manufacturing. This is not limited to the cars, but also includes the phone app, servicing, and customer experience. Other product types (e.g. camera) and companies do this to some extent, but Tesla takes this much further than most. The only products that may exceed what Tesla do are computers, even then it is only Apple that own the software, hardware and customer experience like Tesla do, and Apple do not own the manufacturing.

In future this integration is likely to extend to services, in the Robotaxi and Tesla Network. We already see it in the virtual power plants using powerwalls. Manufacturing is one part of a much bigger Tesla system, perhaps in future embedded in much bigger eco-system of third party applications (e.g. pizza delivery, games, on-the-move productivity software). For Tesla the boundary is not the factory gate.
 
The longer China delays opening up, the weaker their immune systems will get. It's not just covid, it's also influenza, RSV, common colds etc. At some point when they open up it will be like Europeans coming to America. And international students from China going to Australia etc can expect to spend the first few months constantly sick as they play catch up.

I have a feeling that this will turn out to be another great leap forward disaster. They got a few years of low amounts of disease(at a high cost), but in the end they will have to go through a short period with a lot of viruses and weak immunity.

The option is to be closed forever, or at least until technology has found a way to solve immunity for any pathogen.

This might be a bit offtopic, but since Tesla has 50% of their current production in China, I think it's valid to talk about what will happen once they give up zero covid or if they don't. Maybe separate thread soon!?
China only does lockdowns for Covid. The other diseases are still endemic and taken care of by the immune system. It’s only Covid that will have an impact on Chinese people leaving the zero-Covid bubble of China.
 
China only does lockdowns for Covid. The other diseases are still endemic and taken care of by the immune system. It’s only Covid that will have an impact on Chinese people leaving the zero-Covid bubble of China.
His point is probably that the covid measures and lockdowns in China have the side effect of preventing people from catching the usual diseases like the flu, which results in a weakening of immune systems against those diseases as well. I'm not sure what the measures in non-lockdown areas are. If everyone in China has to stay indoors and avoid contact with others if they have a slight cough, then I can see it being a valid point.