Elon Musk Believes Factories Can Improve 10 to 100 Times

Tesla CEO Elon Musk spoke with Automotive News Publisher Jason Stein in the first of a three-part interview.

Elon said that today car factories have a speed of 0.2 meters per second. Walking is five times faster than the speed of an auto assembly line.

Elon wants to be able massively speed up car factories.

Elon talked about
* Tesla having four profitable quarters.
* in the near future, all cars will become autonomous (full self-driving)
* Tesla will have a sustainable advantage against autonomous car makers with better manufacturing

If you optimized every cubic meter of the factory, then you could think of the factory-like a CPU.
You can move things closer together and increase the clock speed to increase auto production throughput.

These optimizations mean a minimum improvement of ten times and up to 100 times.

SOURCES- Auto News interview of Elon Musk
Written by Brian Wang, Nextbigfuture.com (Brian owns shares of Tesla)

32 thoughts on “Elon Musk Believes Factories Can Improve 10 to 100 Times”

  1. can’t disagree that the Economics are hard to justify at the time – but throw in exotic food stuffs availability (even vineyards?); consider near future reduction in the overall farmable land footprint by 15-25% due to sprawl, warming, difficulty in finding labor, and use for animals/ bio-fuels; drastically-cheap urban land now with likely long-term price destruction; special deals on delivering food-stuffs to a large exclusive buyer; and the likely reduction in imports/ exports due to current foreign policy and other uncertainties -and- you may find that there may be a few match-ups for a stacked urban farm through the second/third-tier US cities.

  2. Manufacturing in space makes much more sense when building objects meant to remain in space. Making things that need to be de-orbited causes the problem.

    Knowing Elon Musk as I do (not at all) he probably plans to hyperscale space-based structures that would require a fair amount of space-based manufacturing. If he has the money to do it (he will) then it would make sense to build out before the need arises, just to position himself for the inevitable requirement.

  3. Yes, the competition in the auto industry is a lot fiercer than in space
    industry. Paradoxically, there is far less competition in the luxury segment than in low cost cars, I guess because millionaires don’t
    watch pennies. I will believe Musk claims when he’ll churn out a $10k
    electric car.

  4. I don’t know how dense modern factories are but I’m sure the machinery could be much more closely packed. However i would think that 1 limiting factor would be human safety. Put the machinery too close together and if something breaks down the whole factory may need to be shutdown to make it safe for a human to wander in to fix the issue.

  5. Stacking crops saves land area – but good farmland (or land in an warehouse area near a city) is generally fairly cheap – $10K/acre would be a good price for flat land with decent soil.

    Suppose you built a gigafactory sized ‘farm’ of 2M sqft, stacked 10 high. That’s equivalent to ~450acres – a modest sized farm.

    The Tesla Nevada gigafactory cost $5B to build. An Amazon warehouse of 3.7M sqft (much less dense than a stacked farm building, not stacked nearly as high, and not nearly as automated with 1500 employees) has been estimated to cost $250M. So maybe assume $500M for the automated farm building. That’s $1.1M/acre-equivalent.

    Maybe give it 4 floors, each with 10 high stacks? OK – that’ll need artificial lighting, but ignoring the cost to add more floors, it’s still only down to $275K/acre-equivalent – another factor of 27 to get to parity with farm land. Each yearly payment on a 30yr loan for an acre-equivalent would be enough to buy over an acre of farm land for cash. That’s going to be hard to make up in savings on pesticide and herbicide, assuming you don’t need those. Double-cropping will help some.

  6. The auto OEMs have already done a lot of work at reducing production costs. They reduced manufacturing overhead by outsourcing production of parts and subsystems, and they reduced inventory costs by using just-in-time assembly processes. Don’t know how Tesla will cut costs of current auto production processes by an order of magnitude.

  7. These days every national currency is a non-gold money system. If you want something more futuristic, just about every cryptocurrency is too.

  8. I personally hope for the day of a more ‘artisanal in all things’ world where it is not about factories though these should provide a baseline of essential products. The day of the ‘productive’ and ‘working-with-my-hands’ combo (as a main source of income for 50% of people) are soon over – a re-visit of the build-your-own computer could be a set of super-lab/ studios of ‘maker’ culture. Small inside garden with own custom-food stuffs, mini-maker furniture/ small tools/ household things (fixtures), visual arts studio with advanced additive, etc… The key question: do I spend 5 weekends doing it or do I buy a good ‘industrialized’ version for 2% of my salary instead? Your choice.

  9. Generally if you want to build more things in a smaller space, the equipment you need to do that will have to be more complex, which results in greater expense and less flexibility.

    There’s only so fast things can physically move before you start to have large increases in machine breakdowns and reduction in quality.
    The only way I can see massive speedup in production as Elon describes is if we had self repairing machines and maybe some other kind of nanotech and AI, but we’re a long ways from all of that.

  10. Mining – (mostly) full automation – line-of-sight and tele-operated drill, other impact, detonate, slag removal, in-situ seam and sample gather/ assess set-up, vehicular operations, failed/ underperforming equipment switch-out, and preparation for raw delivery. Assisted by software optimization at production and sync to drill. 95% labour reduction. Humans limited to repair/ start-up Techs, quality control, and site management. Functioning at several Rio Tinto/ BHP sites – australia, south-west, mid-west US. No impact includes horizontal drilling, hydraulic fracturing, etc – don’t know of successful mining outside of oil fracking. Farming – not talking artisanal or ‘local/GMO/low-pesticide for the earth’s sake’ or such – talking average field strains done in low- to no-soil stacked racks with special wavelength lighting and centrally-controlled irrigation/ nutrient feeds. Many, many projects at academic and post-academic facilities – these projects lack the Elon Musk (or other billionaire visionary-personailty)’s intangibles – devoted fellowship, barely-vetted (here’s my money, just associate me with it) angel investing, supply-chain/ regulatory-morass domination, cost-savings-before-woke/greenie-cuteness mentality, and publicity savvy to scale-it up.

  11. Not sure what you mean by no-impact all-in-one Mining operation?

    Kimble Musk is sort of doing the “City Farm” thing, so Elon probably won’t.

    There are foods that don’t transport well, such as old-fashioned strawberries that almost melt in your mouth, sweet corn cooked moments after it is picked, perfectly crisp lettuce, etc – that may meet a high-end/luxury market from urban farming. But I’m fairly certain most plant based foods we eat will be more affected by farm automation, not urban farms.

    Meat, on the other hand, seems ripe for revolution. There’s the impossible burger – but there’s also BlueNalu, growing real fish fillets: https://www.bluenalu.com/pr-121719

  12. Haven’t manufacturer’s been trying to do this for the last century, peoples entire careers are based on this problem. Can he get some speed gains out of this format yeah probably but 100 times yeah nah.

  13. His whole point of view is misleading, He should talk about new manufacturing processes instead of assembly line speed and increasing its speed as if it can be done with a knob. The time spent is not due to the speed of the assembly line, but to speed of assembly. Only idiots will fall for this.

  14. Let me send comments for you.
    10,000 comments cost $1, 100,000 comments cost $5
    You can receive this comment, it means that I sent it successfully.
    If you want to buy, pay 1 or 5 dollars through PayPal, after that,
    Send the following information to my Email:
    Name, email, Url, content.
    I will send the send Log to you by email within three days
    My PayPal & Email: sdzfd@hotmail.com

  15. An increase in speed doesn’t automatically translate into savings. Cost of material wouldn’t be affected. What you might get is a lower cost for labor and capital per unit.

  16. If he can speed things up 100x then the right mix might be speeding things up 10x and lowering cost 10x. Or maybe speeding things up 3x and lowering cost 30x.

    Is factory throughput speed really that important when you can build multiple factories?

  17. Factory design is truly interesting. Many Universities have Mechatronics programs that have now started to bring on advanced process design that is less about assembly lines and more about in-situ vs system flow (move the part around the factory or move the factory around the part). Unfortunately, it starts to get more about re-tooling costs/timelines, degree of customization/modularization(?), and other such unappetizing business/money-flavourings. Mecha post-Docs are pilfered, looted, and lured away from these programs to short-term gigs all over the advanced car/ bus/ train worlds of US/ Germany/ far-east. Almost as lucrative as big Pharma. Of course, the next Big Thing with this kind of automation/production-design is mining and urban-farming. Its a bad decade for traditional mines and farms (and bad last decade for factories) around the world as the equivalent of the Great Shale extraction tech in the US is about to spread to other traditionally blue-collar/small-ownership industries. If Elon wants his next Big Thing, create the perfect City Farm or no-impact all-in-one Mining operation. Factories that make factories – who would have thunk it.

  18. If the de orbiting *vehicle* were made of also valuable material, it could be a part of the value returned. Frozen gas, Argon (environmentally neutral) in particular, can take the hot initial phase, until metal can withstand the temp of slower speed. If we create a small “Bermuda Triangle” of bubbles in the water, a softer but higher speed landing may work, perhaps up to terminal velocity, esp if the *vehicle* is puffy and, perhaps, floats. Try with Gold chunks first. Rotovators may need mass coming down to balance the mass going up, but they seem a generation (machine, not human!) away. Now, the big thing would be to build things such as Starships that can “get down” by design, or things that are used in Space, such as Space Solar, or O’Neill Settlements.

  19. That may be far easier than building a factory in Space, as I suspect the *good* (small!) asteroid will be coated with ice and deorbited by atmos for processing here, at the very beginning. We need to start thinking of a non Gold *money* system, as the price may drop, quite soon.

  20. Getting back down to earth is still the most difficult problem. It’s not going to be easy de-orbiting thousands of Tesla cars per day in good condition.
    A rotovator could reuse most of the deltaV for de-orbiting stuff but terminal velocity from LEO is still high enough to cause major problems.

  21. Nickel is much easier to find. Something like 5% of all space rocks are nickel iron (also have a good amount of cobalt!)

    To start with, just getting anything returned in decent tonnage would be a major advance, so don’t make things difficult for yourself.

  22. OK – but a human-scale two armed robot, including electronics, probably uses less than 5KW power – around 50 cents an hour. And I’d bet its yearly maintenance for 24 hour workdays is a lot less than the cost of medical insurance for 3 shift workers to cover 24 hours.

    Yeah, Musk’s vision of an all-robotic car factory is ahead of its time – though probably if the cars were carefully designed for robotic manufacture, it could be done. I’ve read of a lights-out factory run by FANUC.

  23. Electricity and maintenance are the robot’s salary. I was wondering
    if Elon’s idea of a hyper-automated, almost Berserker factory,
    isn’t too ahead of time, to the point of being uneconomic.

  24. First and foremost, make money. He should go first for palladium,
    gold, and platinum. Then he will be able to buy all the nickel he
    wants, until it costs more than the aforementioned.

  25. “Inexpensive” misses the mark for “bio-robot” humans. Yes, the up-front investment to add a human to an assembly line is pretty low compared to buying a robot – because we gave up buying human workers about 150 years ago. But humans expect to keep getting paid, unlike robots. Clearly if humans were less expensive in general, we would have no robots on any assembly lines.

    Setting aside “wondrous” and “sophisticated” (which manufacturers generally won’t pay for), consider “generalist”. Yes, humans can do a lot of things – but so can robots. There are some things that it is difficult to make a robot do – but there are many things that robots can do much faster and better than humans, sometimes in environments harmful to humans. I think we have to give “generalist” to robots as well, because we use both humans and robots where fixed-function machines don’t make sense.

    The primary human technical advantage over robots is ‘rapid hand-eye coordination’. E.g. humans can more easily deal with floppy stuff like cloth and wires. But robots are being improved in this regard – I’d guess most of this human edge will be gone within a decade, maybe two.

  26. A 10X faster factory does lower production costs, as long as the factory costs less than 10X as much.

  27. He should consider the possibilities of doing this sort of thing in Space. He could get his Nickel there, and refine it there. He can get there. Hmmm . . . (edit: what if an alloy that is only able to be made in 0 g is better than current highly processed Earth steel, in a simple 3D printer?)

  28. Yes. BUT! Remember that Elon has bigger goals — production speed, and particularly rapid-prototype production speed, could be crucial in lunar and martian infrastructure development.

  29. Production speed is important only if it lowers production cost. Let’s hope
    that Musk doesn’t commit again the mistake of not seeing humans as
    wondrous, sophisticated, inexpensive, generalist bio-robots.

Comments are closed.