Sunday Robotics: The household robot we've been waiting for

Stocks A to Z
21

But none of those are things that can’t be easily replicated by other companies. The first is not at all specific to Tesla, as you recognize. The second, vertical integration, is actually quite characteristic of a lot of Chinese manufacturers. The third, manipulation, is a possible example of getting an early lead on something but is not going to be exclusive to Tesla if other companies see that it’s essential and decide to catch up to them.

Again, like EV’s, it certainly seems like humanoid robots would become just as commoditized a product within a few years after the technology was there to serve a mass market.

2 Likes
22

Even Tesla’s EVs still have greater gross margins than most other car companies’ cars of any type. The utility-scale stationary battery products are a great example of something that could get commoditized quickly but Telsa’s margins stay high due to software and other advantages. And iphones have superior margins and network effect compared to other phones due to software and services connected to them. Why can’t one companies’ robots have similar characteristics? How about manufacturing prowess? Even though you can tear apart a product, you still might not be able to replicate the manufacturing trade secrets that created them - for example, Tesla’s new dry electrode battery manufacturing processes.

1 Like
23

But obviously not great enough margins for Tesla to stay really dedicated to the product line. That was captainccs’ point that I was responding to. Tesla has pivoted away from EV manufacturing, and he suggested that was a smart idea because EV’s had become commoditized. Even within a commodity product there can be variations in the profit margins of different companies, but that doesn’t mean that the product isn’t a commodity. So you’ll see different laptop manufacturers or TV manufacturers or microwave manufacturers with different margins, but those products are all highly commoditized.

Because it’s unlikely that humanoid robots will exhibit network effects or provide that kind of software differentiation. Almost by definition, humanoid robots will have extremely similar form factors. The promoted use case for them is to replace boring, repetitive jobs - not the sort of uses that would generate network effects the way that, say, iphones have.

24

I think that assumption is incorrect. One of many examples is in home health care as a substitute for a human. This is absolutely not the same as a microwave oven. It might take a while to get there. But if Tesla is first to the boring, repetitive stuff and then expands to the harder stuff (restocking shelves at Walmart at night, harvesting difficult produce, maintenance on earth, nursing, maintenance in space, etc.) when others compete better in the easier spaces, then they can stay ahead of your commoditization path.

25

Maybe not.

Probably because their autonomy is struggling and a robot is far, far away.

Maybe having some new revenue, even if car manufacturing, has appeal for them, right now?

Tesla new small suv

26

Easy to say. Not easy to do. That’s why legacy automakers can’t made money on EVs. I suggest you read Clayton Christensen’s Innovator’s Dilemma.

AI Overview

Clayton Christensen’s The Innovator’s Dilemma (1997 explains why outstanding companies can do “everything right”—listen to customers, invest in R&D, and maximize profits—yet still lose market leadership to new, disruptive technologies. Established firms often ignore, then succumb to, simpler, cheaper, or inferior technologies that eventually improve and dominate, such as when Blockbuster failed to adapt to streaming.

The Captain

27

Why isn’t it the same as a microwave oven? Commoditization simply means that most of the products share a fairly standard feature set and are mostly indistinguishable among competitors’ offerings. They’ll differ in some modest ways (like unimportant cosmetic features), but for the most part they’re pretty much the same. A microwave is a microwave is a microwave, for the most part. They heat stuff up. Your robot vacuum, your laptop, your computer keyboard - they’re mostly all just doing the same stuff in the same way.

If humanoids reach the point where they can do home health care (and that’s way down the road, they’ll be restocking at Walmart long before that), that’s a pretty commoditized function. Home health care requirements are pretty standard. Personal hygiene (bathing/dressing), light housekeeping, meal preparation, medication reminders, and vital signs monitoring.

For the most likely early uses of humanoid robots, like restocking shelves or working in warehouses), it’s even more standardized. Any humanoid that can perform the job is going to perform it in roughly the same way. It’s a product that’s going to become commoditized.

1 Like
28

I’ve read it, and it’s not relevant here. I didn’t limit myself to legacy competitors. Goods can become quite easily commoditized just by all the new entrants (laptops, for example).

For a brand new product field, like humanoid robots, it’s even less relevant - because there isn’t an incumbent that’s being hindered by the innovator’s dilemma. It’s a new product segment. Or if you think of it as being of a piece with robotics more broadly, then the above point holds - I’m not arguing that existing incumbent robot manufacturers are going to be key players in the humanoid robot game (though they might be), but rather that the product will eventually become commoditized just as much as EV’s are regardless of whether the folks in that commoditized markets are all newcomers or include legacy folks.

29

Eventually it will be commoditized but that is sch a long way off. I think your definition of commoditized is way too broad as is your definition of standardized. You could argue accountants all do the same thing but some get paid Big 4 dollars and some don’t. You could say all lawyers negotiate contracts, argue in court, etc. But obviously there are big differences in what they charge and hopefully big differences in quality that comes with it. Musk wants digital Optimus to handle most of these tasks (via xAI, I think) but will that get “commoditized” within a few years of it first being solved just because we can say they all do X? I don’t think so.

A robot that works with people that need assistance dressing, bathing, and doing other life skills might have a standard set of instructions but the conditions, the patient needs, and the patient behaviors will all be very different. Some robots will do well. Others will not. I expect this will be an extremely hard to problem to solve with a robot. That is what I mean by this is not the same as a microwave. The number of parameters is too great to put into a menu. The software will be critical and if they are distributed to customers via fleet operators, then I can easily see network effects.

Even stocking at a Walmart is, I believe, much less standardized than you seem to think. Products of all sorts of shapes, sizes, rigidities, and weights, sometimes with shelves that need ladders for current humans to reach. This will not be trivial. And again, businesses will have needs that will make SAAS a critical component of the customer experience.

And smart phone all have similar form factors but we seem to be in agreement that iPhones are not commoditized.

30

“Form factor” has nothing to do with it. It’s about creating an impenetrable moat that no one (exaggeration) can cross. Coke is said to have it just because the image of the brand is so universal and so beloved. WalMart is said to have it because of their immense buying power and (more crucially) the distribution network that Sam set up while he was alive, which is unmatched in the retail world.

When it comes to things technical so far, at least, none of those apply. Xerox had a huge brand name, yet when Japanes copiers came in they swept the place. The IBM PC is gone in spite of the parent company’s expertise and history.

Microsoft? Not so much. The developed the OS for a specific environment, then quickly launched customer-ware the worked only on that OS. Once you bought one you had to buy the other. If you had Word, Excel, etc. there were huge switching costs, and nobody wanted to do that. (More to the point, while the IBM PC reigned, business IT guys loved it because “IBM” meant reliability, etc.)

This is “lock in”. Apple has that; switching back and forth from Android to iOS is not like switching from Coke to Pepsi and back. There is a real angst from learning the new system, losing the apps you have paid for, rearranging photo albums, whatever.

In the same way, Warren Buffett’s acquisition of BNSF Railroad has a huge moat, because nobody is going to build another rail line through the suburbs and into the LA Port. It is a “toll bridge” highway that in induplicable anymore, and probably will be forever.

At the moment there is no indication that any one particular robot system (there are dozens!) will have that kind of ‘lock in”, although surely some will be better than others. “Being better” is good, but didn’t save Schlitz from dropping from #1 to #20. Perhaps one of them (or two or more) will find some way to insure that kind of switching cost., although at the moment it’s hard to see how.

But if they do indeed run on “apps” (the nursing home app! The factory app! The cook dinner and clean the kids room app!) maybe there will be a way to make them so unique and so non-interchangeable that that will happen. Again, I don’t see it yet, but nobody sees around every corner, so maybe it will happen.

Being “better” isn’t enough. Lots of companies were “better’ until they weren’t. See: GE, Nike, Blackberry, Yahoo!, PanAm, Ford, etc.

1 Like
31

Yes, of course, eventually everything must come to an end. But I will be very happy if Tesla maintains a competitive advantage as long as GE did.

1 Like
32

IBM’s S/360 S/370 systems ruled the computer roost for two or three decades until they didn’t. These lessons seems to be lost around here, just negative, negative, negative, Tesla can do no good despite it’s track record. I wonder why…?

33

There’s no new product revenue in sight to justify a 300 P/E (never mind profitability).

Tesla bulls:
What new product revenue (product? revenue amount?) are you expecting through 2028?

I asked this in another thread.

Crickets.

The best thesis I can guess is a merger with SpaceX - financial engineering, not product engineering.

34

Because Tesla’s track record has been so mixed.

I don’t think you were around this board at the time, but a few years ago the path to Tesla’s success was going to be its total domination of the automotive market. Tesla was going to grow their automotive sales by a 50% CAGR up to 20 million units per year. This was going to happen because Tesla had manufacturing advantages for EV’s that could not be matched. They had all the elements that would cement their dominance in the field: juggernaut factories that were the “machines that built the machines,” Elon’s first principles approach to manufacturing, the “unboxed” assembly method, the massive lead in battery tech, a “light” approach to different models with minimalist interiors because they’re easier to manufacturer, the absence of dealers and advertising expenses, etc. The were beyond the competition.

But all of that fizzled. Obviously they never got anywhere near the sales numbers. The competition caught up - Tesla’s sales growth in EV’s has stopped (declined!), their margins have shrunk down to ordinary margins, the Cybertruck was a commercial flop, the Semi has been glacial, etc.

For all that The Innovator’s Dilemma gets cited as justification for why Tesla will succeed, they have some attributes of an incumbent that’s been unable to adapt to new business approaches. Newly formed competitors have flooded the market with cheaper models filled with dazzling variety of the bells and whistles that Tesla eschews (because Tesla wants to make the cars that are simplest for it to make, not necessarily what consumers might want).

So, yeah - some of us are rather skeptical that Tesla knows how to defend an early lead in a commercial product once the competition gets to catch up and it becomes less cutting edge and more “ordinary.” Because when that happened in their main business segment, they weren’t able to respond and ended up having to pivot away from the market entirely.

3 Likes
35

He was (see below)…

And don’t forget gigacasting:

And of course, now virtually every EV manufacturer does so-called gigacasting.

1 Like
36

I forgot about gigacasting. And that captainccs was here. And that the word was dreadnought, not juggernaut, to reflect Tesla’s complete reinvention of automotive manufacturing. And presumably about Dre as well.

Anyway, all of those things. The original Tesla story was that they were worth their enormous PE because no one else would be able to catch them - they had such an insurmountable advantage in auto manufacturing because of all their awesome things that they would grow their volume insanely fast, at high margins, to capture 20% of the global market for all autos and hit 20 million units by 2030.

Which, of course, never happened. Which is why those of us who were skeptical of the dreadnought theory of how Tesla would fend off competition when their lead evaporated in a more mature industry are equally skeptical of that lead being maintained in robots (if ever they even get one).

37

This never happened and probably never will happen. But it isn’t because the existing automakers got their act together, they haven’t. What happened is one big thing and one small thing. The big thing is that China out-innovated Tesla rapidly, they produce very good EVs that are substantially less expensive than anything Tesla can produce. Y’all need to travel outside the USA and look around at all the cars on the streets - nearly all the EVs are Chinese models and they are mostly beautiful. And they dominate and will continue to dominate especially in countries with high vehicle purchase taxes. If the tax, is let’s say 40% (VAT plus other assorted taxes), then the tax on a $40,000 BMW is $16,000, but the tax on a roughly equivalent $25,000 BYD is $10,000, just the tax is a $6000 difference, between the price and the tax, there’s a VERY hefty savings of over $20,000!

The small thing that happened is that there was a sudden anti-EV sentiment in the USA from both the right-leaning crowd and the left-leaning crowd. If gasoline prices remain high, that sentiment may soon reverse. But that sentiment caused EV sales growth to slow dramatically. There’s no way Tesla can show a 50% CAGR over 5 years anymore. No way at all.

2 Likes
38

I second this sentiment. If not the the 100% tariff from Biden, China might already be the #1 EV in the US market.

Hawkwin

Who thinks Biden was wrong and that we should have treated China like we treated Japan when we first faced competition from their imports. I am convinced our domestic car market (as well as our consumer) is better for it.

39

If it is first.

I think your path of starting simple and getting more complex over time is logical. Tesla’s approach doesn’t seem to dovetail with that very well, however. Optimus seems to be designed for to do the hard things, when robots still aren’t very good at boring repetitive stuff like stocking shelves. I’m definitely not sure that only Tesla will be the first to make robots that do difficult tasks.

In the 2025 shareholder letter, Andy Jassy mentioned there are now a million robots at Amazon and soon Amazon will have more robot workers than humans. Amazon has three fully integrated robotics divisions (Amazon Robotics, RIVR, and Fauna Robotics), and a couple others than are essentially captive R&D companies like Zoox and Agility Robotics. Plus Amazon has made an enormous investment in a physical AI model with its partner Anthropic.

In other words, Amazon is deploying productive robots right now. It seems logical they will continue to get incrementally better and becoming more general,
while Optimus is still in the blocks.

It is possible that one day when Optimus emerges fully formed and ready to do some work, those boring repetitive tasks will already be performed by a simpler, cheaper robot.

In some ways is this like the self driving car competition. Waymo (and Zoox) are starting from a point of what they think they can do successfully and building on that. Tesla is trying for a general solution that will work everywhere.

40

I think the main thing was the EV tax credit ended. 2025 was a record year for EV sales in the US. As discussed in another thread, used EV prices are hot right now. So I think the demand is there, it is just the price of new EVs is too high.

2 Likes