Tesla Bot Division Will Be the Most Successful Robot Company

Here is why Tesla Bot has an assured level of success.

The Tesla robots do not have to exceed the level of Boston Dynamic robots for agility to win. Boston Dynamics was founded in 1992 and has recently started making about $50 million a year in revenue. Boston Dynamics has 300 top AI researchers. Boston Dynamics was bought by Hyundai for $1.1 billion a few months ago. As cool as the Boston Dynamics robots are the regular consumer is not lining up to buy those robots. They have sold or rented a few thousand robots.

Can Tesla recruit 300 or more additional top AI researchers from Tesla AI day? I would say yes Tesla can. They can pay each top AI person an average of $500,000 per year each or $1 million including extra hardware for a top lab. Tesla has $20 billion in cash and has over $4 billion per year in free cashflow. Tesla and Elon can easily capitalize a larger humanoid research division than Boston Dynamics and Tesla can do it for $300 million per year without buying Boston Dynamics or Sarcos Robotics.

Can Tesla create the real humanoid robots or centaur bots that will be shown in one year at a product launch AI event in one year or worst two years? Yes, Tesla can create working prototypes.

Elon can bring compelling prototype robots on stage in one to two years and then start accepting pre-orders. Cybertruck pre-order levels of success would have Elon surpassing the revenue of Boston Dynamics. 1.5 million $100 deposits would be $150 million.

Just by announcing this vision, Tesla in a few months will have hired a few hundred more AI researchers. Those researchers accelerate the self-driving progress and launch a new humanoid and human scale robot division. Matching existing levels of robotic arm work with navigation will generate billion dollar per year revenue within four years in industries and markets that are $100-200 billion per year opportunities. Tesla can generate revenue from less incremental improvement than other robot companies. Tesla has access to pilot robots at a significant scale in their own factories, construction sites and partner and related businesses. Tesla can bundle partially useful beta robots with their other products to generate utility and sales. These would be useful real-world deployments.

Tesla wins just by hiring a few hundred more AI researchers to accelerate FSD and ensure greater success with FSD.

Tesla can be far more successful at monetizing beta humanoid and human-scale robots than any other robot company. Tesla has a base of real world navigation which enable real-world deployments which is a more useful capability than just agile robots for actual deployments.

Any Tesla bot competitor would need to buy and integrate something like Mobile-eye for vehicle self-driving and Boston Dynamics and add in billions in investment and partner with the most highly capitalized car companies, space companies, solar power and home automation company, battery utility companies and tunneling companies.

Tesla showed off an ATV (All-terrain vehicle) at the Cybertruck event. Tesla can put Vision navigation and the top torso of Tesla bot onto an ATV for certain useful applications.

The 2001 patents for the Segway have expired. Tesla can put the Tesla bot top torse onto balanced wheeled Segway type vehicles.

Tesla already has pure vision navigation AI that works fairly well. The Tesla approach to navigation AI scales down to smaller form factors.

In terms of real-world applications, Elon will apply his five key steps for making things better.

1. Make your requirements less dumb. Question requirements even from very smart people. Whatever requirement you have must have a person responsible for it. If people are questioning the requirement, then you need to go back to a person who knows why the reason it was needed in the first place.

2. Try to delete processes and steps. You want to be forced to add back in at least ten percent of the steps you tried to take out. If you are not forced to add back then you are not working hard enough to remove steps.

In school, we get strait jacketed into not being allowed to question the question or questioner and not being allowed to reject questions and steps.

3. Simplify.

4. Accelerate cycle time.

5. Try to automate it.

Elon has many options for using humanoid or human scale robots in potentially large and lucrative real-world applications.

Boring Company utility tunnels would be tunnels that are 6 inches to 3 feet in diameter.

Elon needs to automate and improve solar power, solar roof and powerwall installations. Tesla will likely get into HVAC with their world-class heat exchangers. This will mean a larger part of the residential and commercial construction markets.

Tesla has options for getting to robotaxi and delivery services. Tesla can add humanoid robots or wheeled robots with the top torso of the Tesla bot. Those can add utility for cybertrucks and cybervans. Tesla will have an advantage attacking the last mile issues for delivery and services.

The robot arms and head can be used to automate the supercharging stations.

SOURCES – Tesla, Brian Wang’s analysis
Written by Brian Wang, Nextbigfuture.com (brian has shares of Tesla)

24 thoughts on “Tesla Bot Division Will Be the Most Successful Robot Company”

  1. Centaur lower body with stumpy legs and wheels on hind hooves, that can stand up like a Boston Dynamic's Handle robot would be an interesting platform. That's sorta Segway-like.

    Reply
  2. Hose down? Just have inbuilt sterilising autoclave function.

    Most of the time it won't activate at the wrong point in the cycle.

    Reply
  3. Brian probably is right that Tesla could invest enough to be able to create, in just a few years, robots that will meet or exceed the physical agility of the robots produced by Boston Dynamics at that date. I believe the rest of this article is a lot less likely.

    Use of robots in such a wide variety of ways would depend on developing what often is called general artificial intelligence. Some small steps have been made in that direction, but I have a strong feeling that our experience in that field is going to be similar to our experience in controlled fusion.

    If applications for robots can be identified that are both fairly narrow in the range of competence needed by the robot, and that robot can be built at a low cost, I think there is a fair chance that Tesla could make a successful business out of it. However, most of the applications Brian mentions do not, I believe, fall into that limited range, so I'm highly skeptical Tesla could be successful at making robots for those applications.

    Common sense is very hard to implement in a computer. I believe it will be quite a while before we have computers with common sense matching that of a six year old child. I believe you would need a robot with more common sense than a six year old child to do many of the tasks Brian envisions for Tesla robots in this article.

    Maybe advances in general artificial intelligence will come sooner than I expect. If so, then more of what Brian projects in the article could come true.

    Reply
  4. That one arm, move and repeat robots do 99% of what we need robots for. The other 1% is hard, so hard we might have to hire a Minimum Wage worker to do it.

    Reply
  5. If Tesla will invest a lot of resources in the project(1000+ workforce) and at least few billions of dollars(in early stages) I am quite confident they will create amazing product. I hope they won't be cheap on this because if they develop something resembling this bot from presentation it could be their best selling product so far. If they figure out mass manufacturing at which they are good and offer good price they will have massive hit. I for example would buy such multifunctional robot instead of a car. AI is progressing exponentially, some proto AGI may be on the table in 2022-23, it would make Tesla Bot extremely usefull.
    I am not Tesla fan and I don't care about Tesla current projects much, but this new project is first that makes me excited and I can't wait to see prototype
    If they pull this off they should become $5T company

    Reply
  6. A Tesla GF building Optimus Humanoid robots could easily sell a million a year @ $100k each. They likely wouldn’t have a marginal cost more than 10% of that with a run of a million. BD Spot costs $70k and isn’t anywhere near as capable.
    That’s $90B/yr in profits from a single GF.

    Reply
  7. I don’t think Boston Dynamics is even competitive with Tesla Robotics. Tesla has focused the whole $700B market cap enterprise on robust solution of computer vision. They’ve spent billions over years developing bespoke FSD NN hardware and Dojo NN training HW from scratch. The structure of their mass produced cars is designed around feeding data to the project.

    They’ve created the brains of mobile robots that can see and manipulate the world around them. Tesla has that in a form that’s already mass produced. Boston Dynamics is an R&D shop that tries to make a few bucks to justify itself. It builds good robot bodies in small quantities with very high costs. It’s much easier for Tesla to mass produce humanoid robot bodies than for BD to scratch the surface of building them brains that see well enough to operate in the real world.

    Reply
  8. I am going to go off topic. I am ecstatic that the EPA is finally banning chlorpyrifos, an organophospate that is making American children lose IQ points, and working memory. https://www.npr.org/2021/08/18/1029144997/epa-will-ban-a-farming-pesticide-linked-to-health-problems-in-children
    There are still other organophosphates they need to ban.

    We also need to ban the use of lead in aviation fuel…especially for crop dusting aircraft. There is lead in the vast majority of baby food because rice is easy to digest and is thus put in most of it. But rice is virtually always planted by aircraft because it is fast, even, and cheap…even compliant with "organic".
    https://www.cnn.com/2019/10/17/health/baby-foods-arsenic-lead-toxic-metals-wellness/index.html

    Reply
  9. Sounds like nonsense and conjecture.

    If Tesla Bot, then automate solar roof install, then take over solar market and HVAC too. Build on Full Self Driving (FSD). There's no autonomous driving regardless of Tesla's naming of such. It seems to be 1, 2, 3, 5 years away. It's coming but, it's not here now and I Musk has called "next year" too often to listen to that as an actual outcome. We got Cyber truck pre-orders. How about the truck?

    Robot in humaniform are not the best or most optimized for most tasks. Not for roofing at all. It's like past Sci-Fi with robots sitting in and driving taxi's. That's not the way we'll do it.

    The idea here is it's all "synergies". It's not. Other than speculation, in the real implementation the technical issue are real and numerous and need to be worked through.

    Yeah it's fun but too starry eyed fandom versus where the real progress will be done.

    Reply
  10. It's true that BD has a lot of experience solving the problems, and their robot parkour demonstrations are pretty impressive.

    But the problems have been getting easier to solve as time goes by, too, due to advances in allied fields. (Some of which Tesla is expert on.) So I think Tesla probably could overtake them in a few years by a massive application of resources combined with their superior corporate culture.

    Speaking of parkour: This just came out.

    https://spectrum.ieee.org/boston-dynamics-atlas-parkour

    Reply
  11. I guess that with 3000 engineers(10x Boston Dynamics) Tesla Bot division could catch up in a year or 2. With the same number it won't be possible, because BD has years of advantage in terms of developing all those thousands of pieces of tech and knowledge which brought their tech to todays level/stage and they are still moving ahead as we speak

    Reply
  12. Calling it now: this will get put in a silicone bodysuit and used as a robo-hooker that gets hosed down with disinfectant between customers. I don't think it is a coincidence that this robot is the exact size and weight of a female human (5'8", 125 lb).

    Reply
  13. It's been announced, but is still an intention on Elon's part.

    Not everything that he does turns to gold (Hyperloop can attest to it), but considering how Tesla's FSD has been turning out he may get something out of this. I remain neutral at this moment.

    Reply

Leave a Comment