Is Tesla a good investment?

[41_swish]

The only thing that I can think of that makes this stock pump year over year is that the market seems to think Tesla is a tech company that just so happens to make cars and not vice versa. If any other car maker had these financials they would not be at this insane of a P.E. ratio.

[ChpBstrd]

The only thing that I can think of that makes this stock pump year over year is that the market seems to think Tesla is a tech company that just so happens to make cars and not vice versa. If any other car maker had these financials they would not be at this insane of a P.E. ratio.

But as the revenue to income map above shows, Tesla is indeed a car company that just so happens to make tech.

[reeshau]

If someone thinks that Tesla isn't a car company, then why are they still building cars?  Why add a 5% margin business to your 85% margin AI business?  (Or, even 15% margin in its heyday)  Apple finally realized this when they launched the iPhone.

I get Elon will do his own thing, but if my investment thesis was all these future things, and I was a serious investor, I would be pounding the table for Elon to push off all the "mechanical" stuff to Foxconn or Magna or somebody.  You can't say they want to own it because they execute at a quality level unattainable elsewhere.

[MustacheAndaHalf]

I think Optimus will take a lot longer to sell than Musk thinks.  There will be rich, early adopters who just want a robot... but then what?  Liberals tend to score higher on "openness to experience", which is what you need to try out an AI robot.  But Musk has aligned himself closely with Trump, alienating himself from that customer base.

There's a book about innovative products called "Crossing the Chasm", which talks about the wide gap between early adopters and the mainstream customer.  I think Optimus could easily fall into that chasm.

[ColoradoTribe]

The only thing that I can think of that makes this stock pump year over year is that the market seems to think Tesla is a tech company that just so happens to make cars and not vice versa. If any other car maker had these financials they would not be at this insane of a P.E. ratio.

But as the revenue to income map above shows, Tesla is indeed a car company that just so happens to make tech.

Tesla is a car company that just happens to...

...be a global leader in AI.
...be a top destination for AI, robotics and engineering grads.
...have the largest and ever growing database of autonomous miles from which to train its AI.
...have more compute power than just just about any company on earth.
...refine litium.
...manufacture electric and soon to be autonomous semi trucks.
...sell billions per year in energy storage products and software like Autobidder.
...create its own sell driving software and hardware.
...manufacture battery cells.
...create new manufacturing techniques like gigacasting that are adopted by competitors.
...builds humanoid robots that are currently working its production line doing useful tasks.
...have relationships with sister companies like xAI, SpaceX, etc. from which to borrow technology and expertise.
...has built-out supply chains to support the above at scale.


Ford and GM are car companies that happen to do none of the above and are soon going to be irrelevant. Once consumers can buy a car that drives you around, all other vehicles become flip phones overnight.

I'm paying $100/month for supervised FSD and with version 13.2.8 I am receiving value for the money. FSD supervised reduces my chances of an accident and makes driving less stressful and tiring and more enjoyable, especially for longer trips.

Anybody who hasn't taken a ride in a Tesla vehicle using version 13 FSD really shouldn't be commenting on Tesla's investment merits.  Anybody, in the face of all the above reality that still wants to think of Tesla as only a car company is being obtuse or is ignorant to reality. Anybody that has been wrong about Tesla's merits as an investment for over a decade really should stop commenting as well IMO, but I know that won't happen.

[ChpBstrd]

I think Optimus will take a lot longer to sell than Musk thinks.  There will be rich, early adopters who just want a robot... but then what?  Liberals tend to score higher on "openness to experience", which is what you need to try out an AI robot.  But Musk has aligned himself closely with Trump, alienating himself from that customer base.

There's a book about innovative products called "Crossing the Chasm", which talks about the wide gap between early adopters and the mainstream customer.  I think Optimus could easily fall into that chasm.

There’s a reason most robots sold today look nothing like a human body. That reason is that we have plenty of human bodies - billions of them, and if you hire them to do a task you can just pay them by the hour rather than making a large upfront investment in purchasing a thing (or various lease agreements in which this cost is factored in).

The bestselling robots today are things like 12 foot arms that make spot welds, forklift-like robots, robots that fit inside pipes, and airborne drones. In other words, things with a radically different form factor than the human body.

Honda’s Asimo humanoid robot from a decade ago was arguably more advanced than Tesla’s remote controlled Optimus, as it could autonomously respond to various cues. Still, it failed to find a market. This suggests to me that Optimus is more about generating sci fi excitement for the stock price than meeting any particular market need. It reminds me of The Boring Company.

[reeshau]

Personally, anybody who can't have a coherent and civil discussion with someone whose opinions differ from their own should not be investing directly themselves.

Myopia is death.

If you can't picture the active threats that could spell the demise of your investment, you are missing risks.  Nothing is perfect, and no company lasts forever, even if its name is still hanging around.  Don't fall prey to the turkey illusion.

[Telecaster]

I think Optimus will take a lot longer to sell than Musk thinks.  There will be rich, early adopters who just want a robot... but then what?  Liberals tend to score higher on "openness to experience", which is what you need to try out an AI robot.  But Musk has aligned himself closely with Trump, alienating himself from that customer base.that chasm.

I think so too, but for different reasons.   Musk said that initially optimus would be targeted towards industrial customers, who would put the robot to work doing boring repetitive tasks on assembly lines and such.  But you don't need a generalized humanoid robot for that type of work.   For example, if the robot is on wheels it is more stable, uses less power, simpler, and cheaper.   And why two arms?  Why not four?  If something is repetitive, you want specialized, not generalized. 

Later, Musk says optimus will be rolled out for home use.  Now, there's where a generalized humanoid robot might make some sense.  You have a bunch of different tasks in an environmental designed for humans.  Folding laundry, washing dishes, etc.  But those tasks aren't so time intensive for me that I need a $30K robot.   I could hire someone to come in a couple times a week and be just as convenient. 

[reeshau]

I think Optimus will take a lot longer to sell than Musk thinks.  There will be rich, early adopters who just want a robot... but then what?  Liberals tend to score higher on "openness to experience", which is what you need to try out an AI robot.  But Musk has aligned himself closely with Trump, alienating himself from that customer base.that chasm.

I think so too, but for different reasons.   Musk said that initially optimus would be targeted towards industrial customers, who would put the robot to work doing boring repetitive tasks on assembly lines and such.  But you don't need a generalized humanoid robot for that type of work.   For example, if the robot is on wheels it is more stable, uses less power, simpler, and cheaper.   And why two arms?  Why not four?  If something is repetitive, you want specialized, not generalized. 

Later, Musk says optimus will be rolled out for home use.  Now, there's where a generalized humanoid robot might make some sense.  You have a bunch of different tasks in an environmental designed for humans.  Folding laundry, washing dishes, etc.  But those tasks aren't so time intensive for me that I need a $30K robot.   I could hire someone to come in a couple times a week and be just as convenient.

It reminds me of when Bill Gates built his megamansion.  One of the things he did was install multiple monitors that could display digital artwork.  They were so high-definition at the time that they cost $80,000 each.

Practical?  No.  Something to say is unusual, and to tinker with?  Yes.

[maizefolk]

Honda’s Asimo humanoid robot from a decade ago was arguably more advanced than Tesla’s remote controlled Optimus, as it could autonomously respond to various cues. Still, it failed to find a market. This suggests to me that Optimus is more about generating sci fi excitement for the stock price than meeting any particular market need. It reminds me of The Boring Company.

In both the comparisons with Waymo, but also in this Asimo comparison it is worth keeping in mind that there are two separate axes: cost and capability.

Asimo cost $2M+ unit. Optimus is targeted at a price point of 20-30k. Figure that there's a level of Musk overpromising in that price point and is Optimus actually launches as a product it comes in at $50k/unit. An equally or even less capable robot at 2.5% of the cost would be viable for a lot more use cases which didn't make sense of Asimo.

The robot taxi shouldn't cost _more_ than a current model 3, it doesn't have any extra sensors or equipment. So that should be less than $35,000/car (that's sale price, some of that is profit, most is cost of goods), while the hardware for each Waymo is estimated to be a couple of hundred thousand dollars. An even modestly less capable automated ride hailing service -- assuming it is "good enough", waiting to see if this is true for Tesla's solution or not -- might be able to undercut Waymo on price or improve on quality of service by having a higher density of cars available if each cost costs 1/7th as much.

TL;DR: consider quality per unit cost, not just absolute quality

[Telecaster]

Anybody who hasn't taken a ride in a Tesla vehicle using version 13 FSD really shouldn't be commenting on Tesla's investment merits.  Anybody, in the face of all the above reality that still wants to think of Tesla as only a car company is being obtuse or is ignorant to reality.

You are confusing the product with the investment case.   Having a great product doesn't guarantee a great investment.   Gigcasting is great, but at some point you still need to sell cars.   Tesla's sales in Q1 were 336,681.  That's the lowest number since Q2 of 2022.   The F-150 Lightning is outselling the Cybertruck.   Venerable GM now has 15% of US EV sales. 

[ColoradoTribe]

I think Optimus will take a lot longer to sell than Musk thinks.  There will be rich, early adopters who just want a robot... but then what?  Liberals tend to score higher on "openness to experience", which is what you need to try out an AI robot.  But Musk has aligned himself closely with Trump, alienating himself from that customer base.that chasm.

I think so too, but for different reasons.   Musk said that initially optimus would be targeted towards industrial customers, who would put the robot to work doing boring repetitive tasks on assembly lines and such.  But you don't need a generalized humanoid robot for that type of work.   For example, if the robot is on wheels it is more stable, uses less power, simpler, and cheaper.   And why two arms?  Why not four?  If something is repetitive, you want specialized, not generalized. 

Later, Musk says optimus will be rolled out for home use.  Now, there's where a generalized humanoid robot might make some sense.  You have a bunch of different tasks in an environmental designed for humans.  Folding laundry, washing dishes, etc.  But those tasks aren't so time intensive for me that I need a $30K robot.   I could hire someone to come in a couple times a week and be just as convenient.

Presumably, if their is a job on a factory line being done by a human currently, instead of a specialized robot, to complete a repetitive task (i.e., refilling or stocking components into a larger machine) the company would have already replaced the human. Replacing those humans on the line with a humanoid robot is a no brainer. The humanoid robot doesn't need sleep, doesn't receive pay or benefits, doesn't get sick, or need periodic replacement with a new human that has to be be trained. Robots don't need a break room or take breaks and don't need HR meetings. Additional training is uploaded in seconds. The cost of of a $30k humanoid robot working a factory line would be recouped in under a year in savings. According to Grok a conservative estimate for the number of factory workers engaged in basic manufacturing world  wide is 300 million.

Then there is a whole slew of new labor uses we will find for humanoid robots once labor becomes so cheap compared to human labor. Would an apartment building or hotel in the city hire a doorman to man their door/lobby 24/7? Only in the most upscale apartments now. But, what if an Optimus could do the job of opening doors for a fraction of the cost over a five year period. Conservatively, the human doorman costs over $750k for 5 years by way of uniforms, training, paid holidays, sick days, salary, and a basic healthcare plan. The Optimus robot will cost under $50k for the same period factoring initial purchase cost, electricity cost, and some maintenance and repair costs. That cost will go lower over time with economies of scale lowering the costs of the robots. This is just one of 100s of example where a lower cost for labor will create new jobs and uses that can only be filled by a humanoid robot and its cost structure.

There's a lot of narrow minded thinking that cannot or refuses to grasp the level of disruption about to be unleashed on the world.

[GuitarStv]

I think Optimus will take a lot longer to sell than Musk thinks.  There will be rich, early adopters who just want a robot... but then what?  Liberals tend to score higher on "openness to experience", which is what you need to try out an AI robot.  But Musk has aligned himself closely with Trump, alienating himself from that customer base.that chasm.

I think so too, but for different reasons.   Musk said that initially optimus would be targeted towards industrial customers, who would put the robot to work doing boring repetitive tasks on assembly lines and such.  But you don't need a generalized humanoid robot for that type of work.   For example, if the robot is on wheels it is more stable, uses less power, simpler, and cheaper.   And why two arms?  Why not four?  If something is repetitive, you want specialized, not generalized. 

Later, Musk says optimus will be rolled out for home use.  Now, there's where a generalized humanoid robot might make some sense.  You have a bunch of different tasks in an environmental designed for humans.  Folding laundry, washing dishes, etc.  But those tasks aren't so time intensive for me that I need a $30K robot.   I could hire someone to come in a couple times a week and be just as convenient.

Presumably, if their is a job on a factory line being done by a human currently, instead of a specialized robot, to complete a repetitive task (i.e., refilling or stocking components into a larger machine) the company would have already replaced the human. Replacing those humans on the line with a humanoid robot is a no brainer. The humanoid robot doesn't need sleep, doesn't receive pay or benefits, doesn't get sick, or need periodic replacement with a new human that has to be be trained. Robots don't need a break room or take breaks and don't need HR meetings. Additional training is uploaded in seconds. The cost of of a $30k humanoid robot working a factory line would be recouped in under a year in savings. According to Grok a conservative estimate for the number of factory workers engaged in basic manufacturing world  wide is 300 million.

Then there is a whole slew of new labor uses we will find for humanoid robots once labor becomes so cheap compared to human labor. Would an apartment building or hotel in the city hire a doorman to man their door/lobby 24/7? Only in the most upscale apartments now. But, what if an Optimus could do the job of opening doors for a fraction of the cost over a five year period. Conservatively, the human doorman costs over $750k for 5 years by way of uniforms, training, paid holidays, sick days, salary, and a basic healthcare plan. The Optimus robot will cost under $50k for the same period factoring initial purchase cost, electricity cost, and some maintenance and repair costs. That cost will go lower over time with economies of scale lowering the costs of the robots. This is just one of 100s of example where a lower cost for labor will create new jobs and uses that can only be filled by a humanoid robot and its cost structure.

There's a lot of narrow minded thinking that cannot or refuses to grasp the level of disruption about to be unleashed on the world.

See, your doorman example is a great one against use of robots.

It costs a couple hundred dollars to install an automatic door opener with a scanning keycard.  Let's get crazy and say 5,000$ all in.  Automating this task can be done much, much, much easier than having a doorman.

But why do doormen exist at all?  It's not really to open doors.  They're there partly as a 'look how rich we are, we can hire a dude just to stand there and look pretty' flex, partly do to random general tasks - call a taxi, carry boxes, act as security (checking if people are tailgating when residents enter), sign for/accept packages for guests, answer information queries, general cleaning if the weather outside is poor and the entrance-way gets wet and needs to be mopped.  I don't think Optimus would fill this use case very well at all.

[ColoradoTribe]

Anybody who hasn't taken a ride in a Tesla vehicle using version 13 FSD really shouldn't be commenting on Tesla's investment merits.  Anybody, in the face of all the above reality that still wants to think of Tesla as only a car company is being obtuse or is ignorant to reality.

You are confusing the product with the investment case.   Having a great product doesn't guarantee a great investment.   Gigcasting is great, but at some point you still need to sell cars.   Tesla's sales in Q1 were 336,681.  That's the lowest number since Q2 of 2022.   The F-150 Lightning is outselling the Cybertruck.   Venerable GM now has 15% of US EV sales.

Cybertruck outsold F-150 in 2024. More importantly, Tesla made a profit on its Cybertruck sales and Ford lost billions on its EVs including a loss on every F-150 Lightning sold. Gigacastings are not a consumer product, but hey are just one example of what makes Tesla vehicles a superior product, innovative and Tesla one of only two profitable EV makers in the world. Tesla investment case is no longer centered on vehicles sales numbers. I have accepted that as an investor and explained in great detail above multiple times how Tesla goes beyond vehicle sales. Energy revenue alone will outgain Tesla vehicle sales in a couple years with new megapack factories opening in Austin and Shanghai, followed by FSD, robotaxi revenue, and humanoid robots.

FSD is a product. One with a huge TAM and near 100% profit margin. It can be licensed to other vehicle manufacturers, sold as a subscription, or tacked on as as a one-time expense to the price of a vehicle. Sounds like you are the one failing to understand the disruption, the market, and the investment opportunity. I don't really care other than the flow of misinformation that spews from the bear echo chamber that this thread has become is a disservice to honestly inquisitive folks considering an investment.

[ColoradoTribe]

I think Optimus will take a lot longer to sell than Musk thinks.  There will be rich, early adopters who just want a robot... but then what?  Liberals tend to score higher on "openness to experience", which is what you need to try out an AI robot.  But Musk has aligned himself closely with Trump, alienating himself from that customer base.that chasm.

I think so too, but for different reasons.   Musk said that initially optimus would be targeted towards industrial customers, who would put the robot to work doing boring repetitive tasks on assembly lines and such.  But you don't need a generalized humanoid robot for that type of work.   For example, if the robot is on wheels it is more stable, uses less power, simpler, and cheaper.   And why two arms?  Why not four?  If something is repetitive, you want specialized, not generalized. 

Later, Musk says optimus will be rolled out for home use.  Now, there's where a generalized humanoid robot might make some sense.  You have a bunch of different tasks in an environmental designed for humans.  Folding laundry, washing dishes, etc.  But those tasks aren't so time intensive for me that I need a $30K robot.   I could hire someone to come in a couple times a week and be just as convenient.

Presumably, if their is a job on a factory line being done by a human currently, instead of a specialized robot, to complete a repetitive task (i.e., refilling or stocking components into a larger machine) the company would have already replaced the human. Replacing those humans on the line with a humanoid robot is a no brainer. The humanoid robot doesn't need sleep, doesn't receive pay or benefits, doesn't get sick, or need periodic replacement with a new human that has to be be trained. Robots don't need a break room or take breaks and don't need HR meetings. Additional training is uploaded in seconds. The cost of of a $30k humanoid robot working a factory line would be recouped in under a year in savings. According to Grok a conservative estimate for the number of factory workers engaged in basic manufacturing world  wide is 300 million.

Then there is a whole slew of new labor uses we will find for humanoid robots once labor becomes so cheap compared to human labor. Would an apartment building or hotel in the city hire a doorman to man their door/lobby 24/7? Only in the most upscale apartments now. But, what if an Optimus could do the job of opening doors for a fraction of the cost over a five year period. Conservatively, the human doorman costs over $750k for 5 years by way of uniforms, training, paid holidays, sick days, salary, and a basic healthcare plan. The Optimus robot will cost under $50k for the same period factoring initial purchase cost, electricity cost, and some maintenance and repair costs. That cost will go lower over time with economies of scale lowering the costs of the robots. This is just one of 100s of example where a lower cost for labor will create new jobs and uses that can only be filled by a humanoid robot and its cost structure.

There's a lot of narrow minded thinking that cannot or refuses to grasp the level of disruption about to be unleashed on the world.

See, your doorman example is a great one against use of robots.

It costs a couple hundred dollars to install an automatic door opener with a scanning keycard.  Let's get crazy and say 5,000$ all in.  Automating this task can be done much, much, much easier than having a doorman.

But why do doormen exist at all?  It's not really to open doors.  They're there partly as a 'look how rich we are, we can hire a dude just to stand there and look pretty' flex, partly do to random general tasks - call a taxi, carry boxes, act as security (checking if people are tailgating when residents enter), sign for/accept packages for guests, answer information queries, general cleaning if the weather outside is poor and the entrance-way gets wet and needs to be mopped.  I don't think Optimus would fill this use case very well at all.

Exactly, lots of hotel chains and apartment complexes WILL buy an Optimus to create that experience for their tenants because they can now afford to do it and it will be a differentiator. An automatic door opener is not that same as having a door opened for you. Or having an umbrella held over your head, a taxi hailed, or asking for directions or where the ice machine is located. A simple automatic door opener on a motion sensor does give you any of that.

This is just one example! Cities will buy these in bulk to remove graffiti, pick up litter from parks, man vehicle gates and welcome kiosks. Caterers will have humanoid staffs delivering drinks and bacon-wrapped shrimp. Sky scrapper window cleaners. Gas station attendants to check tire pressure and wipe windows while you charge. Building custodial staff to mop and wax floors. Millions upon millions of existing menial labor jobs and jobs that will become economical to staff once the cost of the labor is reduced by a full order of magnitude. 

[joemandadman189]

Anybody who hasn't taken a ride in a Tesla vehicle using version 13 FSD really shouldn't be commenting on Tesla's investment merits.

I took a Cyber truck on a test drive recently, i have wanted one for a long time and put a deposit down but never purchased one as they were $80k after the foundation models came out, much more than the originally advertised price. I do own a 2022 ram 1500, compared to the cyber truck its an antique, the air ride is sooooo much nicer. The technology is crazy amazing. We demo-ed FSD, it was awesome, kept pace with traffic, slowed down to avoid jay-walker, it even showed the person on the screen. It actually startled me when it made a lane change, i grabbed the steering wheel to get back in my lane and FSD turned off automatically. it can park itself, and i saw another person use summon at a grocery store which was pretty wild. If i wanted/needed a truck today i would get one, hands down, but i already own my ram outright. Maybe a used one in a few years

i know this may surprise the mustashians here but while at the dealership getting an oil change a saw that a new 2025 Ram with leather seats is now $80k, same price as a Cyber truck. Granted the truck was marked down about $15k to $65k but this astronomical price point isnt out of the range of other manufacturers for similar vehicles.

[MinorMiner]

Anybody who hasn't taken a ride in a Tesla vehicle using version 13 FSD really shouldn't be commenting on Tesla's investment merits.  Anybody, in the face of all the above reality that still wants to think of Tesla as only a car company is being obtuse or is ignorant to reality. Anybody that has been wrong about Tesla's merits as an investment for over a decade really should stop commenting as well IMO, but I know that won't happen.

This is the key thing to realize here. The guys on here with positions in Tesla are trading on material non public information. We are all driving the software and seeing what it is capable of and how fast it is developing. Only a very small fraction of the American population has had a chance to do this (Maybe 2%?) I would take this first hand information over 1000 stories from the media.

@ColoradoTribe : When do you think V14 rolls out to the public?

[dragoncar]

If someone thinks that Tesla isn't a car company, then why are they still building cars?  Why add a 5% margin business to your 85% margin AI business?  (Or, even 15% margin in its heyday)  Apple finally realized this when they launched the iPhone.

I get Elon will do his own thing, but if my investment thesis was all these future things, and I was a serious investor, I would be pounding the table for Elon to push off all the "mechanical" stuff to Foxconn or Magna or somebody.  You can't say they want to own it because they execute at a quality level unattainable elsewhere.

Cars on the road provides training data.   Insidious telematics, too.

Liberals tend to score higher on "openness to experience", which is what you need to try out an AI robot.

Well conservatives tend to score higher on "openness to slavery" so maybe they'd like a robot slave

[reeshau]

If someone thinks that Tesla isn't a car company, then why are they still building cars?  Why add a 5% margin business to your 85% margin AI business?  (Or, even 15% margin in its heyday)  Apple finally realized this when they launched the iPhone.

I get Elon will do his own thing, but if my investment thesis was all these future things, and I was a serious investor, I would be pounding the table for Elon to push off all the "mechanical" stuff to Foxconn or Magna or somebody.  You can't say they want to own it because they execute at a quality level unattainable elsewhere.

Cars on the road provides training data.   Insidious telematics, too.

I'm not talking software. I'm talking about fit and finish.  Panel gaps, trim falling off, that sort of thing.  Magna is not well-known outside the industry,  but is the Foxconn of cars.  And, Foxconn itself wants to get into building cars.  The maximizing profit, nimble answer would be to outsource that kind of work.

[ColoradoTribe]

Anybody who hasn't taken a ride in a Tesla vehicle using version 13 FSD really shouldn't be commenting on Tesla's investment merits.

I took a Cyber truck on a test drive recently, i have wanted one for a long time and put a deposit down but never purchased one as they were $80k after the foundation models came out, much more than the originally advertised price. I do own a 2022 ram 1500, compared to the cyber truck its an antique, the air ride is sooooo much nicer. The technology is crazy amazing. We demo-ed FSD, it was awesome, kept pace with traffic, slowed down to avoid jay-walker, it even showed the person on the screen. It actually startled me when it made a lane change, i grabbed the steering wheel to get back in my lane and FSD turned off automatically. it can park itself, and i saw another person use summon at a grocery store which was pretty wild. If i wanted/needed a truck today i would get one, hands down, but i already own my ram outright. Maybe a used one in a few years

i know this may surprise the mustashians here but while at the dealership getting an oil change a saw that a new 2025 Ram with leather seats is now $80k, same price as a Cyber truck. Granted the truck was marked down about $15k to $65k but this astronomical price point isnt out of the range of other manufacturers for similar vehicles.

Good point, most folks think Tesla vehicles are at the high end of the price range, when the average selling prices are right around the current averages for new cars and trucks. New, cheaper variants based on 3 and Y coming out this year will do a lot to expand the addressable market.

Buying a new car is tough for the frugal, FIRE types. I was a Tesla investor for over a decade before I allowed myself to buy a Model Y. Though not cheap, it's the best value for the money of any car I've ever owned. Not even close.

[ColoradoTribe]

Anybody who hasn't taken a ride in a Tesla vehicle using version 13 FSD really shouldn't be commenting on Tesla's investment merits.  Anybody, in the face of all the above reality that still wants to think of Tesla as only a car company is being obtuse or is ignorant to reality. Anybody that has been wrong about Tesla's merits as an investment for over a decade really should stop commenting as well IMO, but I know that won't happen.

This is the key thing to realize here. The guys on here with positions in Tesla are trading on material non public information. We are all driving the software and seeing what it is capable of and how fast it is developing. Only a very small fraction of the American population has had a chance to do this (Maybe 2%?) I would take this first hand information over 1000 stories from the media.

@ColoradoTribe : When do you think V14 rolls out to the public?

Great point and so true. Owning the vehicle and driving on FSD does feel like insider trading sometimes. Especially with so much FUD around the company in the media. Almost without exception, the folks going on about panel gaps and FSD killing people, have never owned a Tesla or ridden around using FSD. I guess we should be thanking them. I bought a lot of discounted shares between 2013 and 2019 based on all that FUD.

[MustacheAndaHalf]

I think Optimus will take a lot longer to sell than Musk thinks.  There will be rich, early adopters who just want a robot... but then what?  Liberals tend to score higher on "openness to experience", which is what you need to try out an AI robot.  But Musk has aligned himself closely with Trump, alienating himself from that customer base.

There's a book about innovative products called "Crossing the Chasm", which talks about the wide gap between early adopters and the mainstream customer.  I think Optimus could easily fall into that chasm.

There’s a reason most robots sold today look nothing like a human body. That reason is that we have plenty of human bodies - billions of them, and if you hire them to do a task you can just pay them by the hour rather than making a large upfront investment in purchasing a thing (or various lease agreements in which this cost is factored in).

The bestselling robots today are things like 12 foot arms that make spot welds, forklift-like robots, robots that fit inside pipes, and airborne drones. In other words, things with a radically different form factor than the human body.

Honda’s Asimo humanoid robot from a decade ago was arguably more advanced than Tesla’s remote controlled Optimus, as it could autonomously respond to various cues. Still, it failed to find a market. This suggests to me that Optimus is more about generating sci fi excitement for the stock price than meeting any particular market need. It reminds me of The Boring Company.

Agreed - robots are already present in manufacturing and other places where a humanoid robot wouldn't be efficient.  Tesla's general robot will struggle to compete with specialized industrial robots.

According to this Tesla stock analyst, Optimus robots are already used by Tesla in its manufacturing.  Reading between the lines, it sounds like this analyst thinks Tesla could build up AI capability through experience with Optimus - perhaps helping some other products.  I don't see any discussion of humanoid robot sales.
https://finance.yahoo.com/news/analyst-revisits-tesla-stock-price-114110040.html

[Telecaster]

Presumably, if their is a job on a factory line being done by a human currently, instead of a specialized robot, to complete a repetitive task (i.e., refilling or stocking components into a larger machine) the company would have already replaced the human. Replacing those humans on the line with a humanoid robot is a no brainer.

The first part is correct.   The reason they haven't been replaced is there are lots of jobs robots aren't good at.  Robots are bad at jobs requiring tactile information or fine motor control, for example.   This is why shoe making still requires lots of human labor for instance.  The second part is wrong.  If you could figure out how a robot could make shoes, it wouldn't have to be humanoid.   It could look like anything you want.   One thing you definitely would not do is make it bipedal because of the higher operational costs and shorter duty cycle.    Turns out it actually is a brainer. 

Cybertruck outsold F-150 in 2024. More importantly, Tesla made a profit on its Cybertruck sales and Ford lost billions on its EVs including a loss on every F-150 Lightning sold. Gigacastings are not a consumer product, but hey are just one example of what makes Tesla vehicles a superior product, innovative and Tesla one of only two profitable EV makers in the world.

Hold up there.   This is an investing discussion.  Let's use terms correctly.    Tesla had positive gross margins on the Cybertruck.    You know what is not included in the gross margin?   The R&D, design, and sales costs, and the cost of building the Giga factory.  Tesla is not making a profit on the Cybertruck.  Which by the way, is not necessarily a bad thing.  You wouldn't reasonably assume a new product line is profitable immediately.    Tesla itself wasn't profitable for years.   

Anyway my point wasn't about comparing Ford and Tesla.  It was about Tesla losing both sales and market share in the US.   Losing margin too, by the way. 

But let's make the bull case for Tesla.   Tesla's EPS has grown on average of 14.20% over the last three years.   If it were to continue that same growth rate for the next 12 years it would have an EPS of $9.79 by 2037.   If the stock price remained exactly the same for the next 12 years, the P/E would be 35.   Which is currently higher than Amazon, Apple, Microsoft, Meta, and Google.   That's a heavy lift.  And of course, the EPS is falling dramatically.   We have to use some very generous assumptions to come up with a scenario where Tesla is still very expensive in 12 years.

The part that keeps getting lost is you can argue that Tesla is a perfectly wonderful company.   But that doesn't make it a perfectly wonderful investment.    That's the question in the title of this thread. 

[ColoradoTribe]

I think Optimus will take a lot longer to sell than Musk thinks.  There will be rich, early adopters who just want a robot... but then what?  Liberals tend to score higher on "openness to experience", which is what you need to try out an AI robot.  But Musk has aligned himself closely with Trump, alienating himself from that customer base.

There's a book about innovative products called "Crossing the Chasm", which talks about the wide gap between early adopters and the mainstream customer.  I think Optimus could easily fall into that chasm.

There’s a reason most robots sold today look nothing like a human body. That reason is that we have plenty of human bodies - billions of them, and if you hire them to do a task you can just pay them by the hour rather than making a large upfront investment in purchasing a thing (or various lease agreements in which this cost is factored in).

The bestselling robots today are things like 12 foot arms that make spot welds, forklift-like robots, robots that fit inside pipes, and airborne drones. In other words, things with a radically different form factor than the human body.

Honda’s Asimo humanoid robot from a decade ago was arguably more advanced than Tesla’s remote controlled Optimus, as it could autonomously respond to various cues. Still, it failed to find a market. This suggests to me that Optimus is more about generating sci fi excitement for the stock price than meeting any particular market need. It reminds me of The Boring Company.

Agreed - robots are already present in manufacturing and other places where a humanoid robot wouldn't be efficient.  Tesla's general robot will struggle to compete with specialized industrial robots.

According to this Tesla stock analyst, Optimus robots are already used by Tesla in its manufacturing.  Reading between the lines, it sounds like this analyst thinks Tesla could build up AI capability through experience with Optimus - perhaps helping some other products.  I don't see any discussion of humanoid robot sales.
https://finance.yahoo.com/news/analyst-revisits-tesla-stock-price-114110040.html

Did you not read the posts up thread. There are currently in the world 300 million humans working in factories that do some form of manufacturing. These are human manufacturing jobs where there is not a non-humanoid robot alternative. It is pretty obvious that the bulk of these 300 million human factory workers could be replaced by humanoid robots and the robot replacement would pay for itself in short order. If you have a line job being done by humans in three 8-hours shifts per day. Even if the worker is only earning $10/hr, that's a labor cost of $87,600/year to staff that job with humans. Optimus will cost around $30k to start and go down in cost from there with economies of scale. There are literally millions of factory workers whose jobs can be done more safely and cheaply by a humanoid robot/AI.

This is before we factor in all the other use cases, some of which I listed above, or use in the home for wealthier households. There's also the jobs that will become economical to staff or creat once the cost of human form factor labor drops an order of magnitude.

Catering Staff
Door man
Gardener
Custodial Jobs
Sky scraper window cleaner
Bar tender
Help Kiosk
Litter Control
Bell Hop
Landscaping
Driving Range Attendant
Crossing Guard
Dog walker
Cafeteria Line Worker
Grocery Bagger
Laundry Room Worker (hotel)
Pool Attendant
Snack Kiosk
Bus Boy
Drink Runner
Stocking Shelves
Order Fulfillment
Security
Landmine Removal
Coat Checker
Bathroom Attendant
Recyling Line Sorter
Dishwasher

[ColoradoTribe]

Presumably, if their is a job on a factory line being done by a human currently, instead of a specialized robot, to complete a repetitive task (i.e., refilling or stocking components into a larger machine) the company would have already replaced the human. Replacing those humans on the line with a humanoid robot is a no brainer.

The first part is correct.   The reason they haven't been replaced is there are lots of jobs robots aren't good at.  Robots are bad at jobs requiring tactile information or fine motor control, for example.   This is why shoe making still requires lots of human labor for instance.  The second part is wrong.  If you could figure out how a robot could make shoes, it wouldn't have to be humanoid.   It could look like anything you want.   One thing you definitely would not do is make it bipedal because of the higher operational costs and shorter duty cycle.    Turns out it actually is a brainer. 

Cybertruck outsold F-150 in 2024. More importantly, Tesla made a profit on its Cybertruck sales and Ford lost billions on its EVs including a loss on every F-150 Lightning sold. Gigacastings are not a consumer product, but hey are just one example of what makes Tesla vehicles a superior product, innovative and Tesla one of only two profitable EV makers in the world.

Hold up there.   This is an investing discussion.  Let's use terms correctly.    Tesla had positive gross margins on the Cybertruck.    You know what is not included in the gross margin?   The R&D, design, and sales costs, and the cost of building the Giga factory.  Tesla is not making a profit on the Cybertruck.  Which by the way, is not necessarily a bad thing.  You wouldn't reasonably assume a new product line is profitable immediately.    Tesla itself wasn't profitable for years.   

Anyway my point wasn't about comparing Ford and Tesla.  It was about Tesla losing both sales and market share in the US.   Losing margin too, by the way. 

But let's make the bull case for Tesla.   Tesla's EPS has grown on average of 14.20% over the last three years.   If it were to continue that same growth rate for the next 12 years it would have an EPS of $9.79 by 2037.   If the stock price remained exactly the same for the next 12 years, the P/E would be 35.   Which is currently higher than Amazon, Apple, Microsoft, Meta, and Google.   That's a heavy lift.  And of course, the EPS is falling dramatically.   We have to use some very generous assumptions to come up with a scenario where Tesla is still very expensive in 12 years.

The part that keeps getting lost is you can argue that Tesla is a perfectly wonderful company.   But that doesn't make it a perfectly wonderful investment.    That's the question in the title of this thread.

I have seven figure gains based on my Tesla investment, so I think I can say it's a "wonderful" investment. I built a custom house using mostly gains from TSLA. So, for me, the title of the thread has been answered emphatically. I suppose the open question is will Tesla continue to be a good investment going forward and have its 3rd 10X run. I've laid out my case in great detail. You can choose to discount the arguments and facts I put forward, but from my perspective there's very little of merit that the bears have brought to this discussion or that would leave me to believe FSD, Optimus and AI won't usher in another wave of revenue growth for Tesla.

[waltworks]

Even if you can make a humanoid robot that can bag groceries, it's going to cost a heck of a lot more than a teenager, though, right? I build and maintain machines that are many orders of magnitude simpler (bicycles) and they still require quite a bit of work to keep running well. I can't even imagine the maintenance challenges involved with, say, a robot hand capable of handling a ripe tomato or something. Let alone the rest of the robot.

From an engineering standpoint, the likelihood seems really high that a hyper-complex device interacting with humans IRL is going to get damaged a lot. And in an industrial setting that's more controlled, there's not much reason for the humanoid form as others have already pointed out.

It's probably irrelevant regardless for quite a while. There's no evidence that Tesla has anything approaching a robot that can bag groceries, teleoperated drink-service notwithstanding.

[maizefolk]

Manufacturing volume tend to go along with more optimized/lower maintenance designs. I've worked with research instruments that are essentially had build by skilled techs, limited run goods (say a thermocycler, or a precision four row air planter), and super widely produced items (cars, laptop computers, phones). Bigger production runs both means it makes sense to spend more time identifying and reducing potential failure/maintenance issues and that, with large number of devices out in the world, over time, you have more chances to identify what the actual failure and maintenance issues are that need to be addressed.

In the hypothetical world were any company had a low-ish cost robot capable of bagging groceries AND they were able to sell hundreds of thousands of those robots, I suspect that over time the maintenance requirements would look more like that of a car* than of a piece of research equipment or industrial automation. I certainly hope it'd look more like a car than like a cell phone/laptop where if anything serious breaks many people just throw it away and replace it.

*Regular periodic maintenance, occasional expensive repairs at a local shop with specialized mechanics when something breaks/wears out or when it experienced the robotic equivalent of being rear-ended or sideswiped.

[ChpBstrd]

Manufacturing volume tend to go along with more optimized/lower maintenance designs. I've worked with research instruments that are essentially had build by skilled techs, limited run goods (say a thermocycler, or a precision four row air planter), and super widely produced items (cars, laptop computers, phones). Bigger production runs both means it makes sense to spend more time identifying and reducing potential failure/maintenance issues and that, with large number of devices out in the world, over time, you have more chances to identify what the actual failure and maintenance issues are that need to be addressed.

In the hypothetical world were any company had a low-ish cost robot capable of bagging groceries AND they were able to sell hundreds of thousands of those robots, I suspect that over time the maintenance requirements would look more like that of a car* than of a piece of research equipment or industrial automation. I certainly hope it'd look more like a car than like a cell phone/laptop where if anything serious breaks many people just throw it away and replace it.

*Regular periodic maintenance, occasional expensive repairs at a local shop with specialized mechanics when something breaks/wears out or when it experienced the robotic equivalent of being rear-ended or sideswiped.

You're getting into very important points. The thing that seems to sell cars is looks that give an impression of coolness, sexiness, strength, etc. Ask the people signing up for 7 year leases on pretty trucks. They thought it looked cool. They liked the facade redesign. They thought it would get them respect.

Will something similar apply to $30,000 AI-driven robot-people marketed to the masses? Or will the "general" robot market be dominated by the question "is it profitable to use?" the way the specialist robot market already is?

I suppose the case for Optimus rests on the assumption that people will buy humanoid robots for the same reasons they buy brand new cars. A specialist robot like the arms that weld car bodies together or the pucks that vacuums floors will always be cheaper to do any specific function, but the AI integration would supposedly make it possible for a general robot to go from welding to bagging groceries to helping nursing home patients to vacuuming the cat hair.

Then we must assume that all those joints, servos, motors, cameras, sensors, processors, batteries, hydraulic systems, etc. are durable and low-maintenance enough so that the robot can earn back its cost over a period of perhaps decades. And the long-term energy bill needs to be evaluated too. What are we looking at? 500 watts just to stand there? These machines would need to be produced to an extraordinary quality level to financially justify themselves... unless they are bought as vanity products, in which case they can be as disposable as a modern Audi SUV or set of chrome pretty truck rims.

Yet, I'm also missing their surveillance value. The reason you can buy a 65" 4K television for under $400 is because the TV is spying on you, sending advertisers information about what you're watching so that ads can be targeted to your specific demographics and interests. Perhaps your robot will be similarly discounted, but will steer you toward purchases, advertisements, fake news sources, or subscription upgrades. Perhaps we'll accept ad-subsidized robots in our lives the same way we accept it from our TVs and phones?

[maizefolk]

I suppose the case for Optimus rests on the assumption that people will buy humanoid robots for the same reasons they buy brand new cars. A specialist robot like the arms that weld car bodies together or the pucks that vacuums floors will always be cheaper to do any specific function, but the AI integration would supposedly make it possible for a general robot to go from welding to bagging groceries to helping nursing home patients to vacuuming the cat hair.

I don't know that I necessarily agree with your assertion in bold (my emphasis).

Economies of scale in manufacturing are real and dramatic and they become more of a factor the more complex the thing you're manufacturing is. It is the reason you see general purpose computers like raspberry pi's or arduino's used in a lot of situations where their capabilities are dramatic overkill vs a simpler custom part. It can be more cost effect to buy a more complex and generally capable computer which is designed once and manufactured as massive scale than to design and produce a much more limited production run of a simpler dedicated component for that use case.

So while I don't know which way the economics would actually work out in this particular scenario (and I also don't know if the generalizable humanoid robot is going to be possible to create and/or manufacture at scale), I think it is at least conceptually possible for a more complex general purpose robot that can do 500 different tasks and is manufactured on a massive scale might end up being cheaper for end users than designing 500 different simpler specialized robots, each with much more limited production runs where the costs of both R&D and capital investments like building assembly lines need to be spread out over fewer units.

[GuitarStv]

I'd be interested to see what operating costs are for one of these robots.  Just maintaining my bicycle, I'm constantly amazed at how much grime gets into spots you don't want it and the damage that this does to the bike as it's used.  If the robots are being used in relatively clean environments (grocery store) maybe this wouldn't be an issue . . . but most manufacturing environments that I've been in get pretty dirty on a regular basis.

[EscapeVelocity2020]

Just a couple random thoughts -

1.  Tesla kills someone during the limited Austin rollout of FSD (not far fetched given their relatively recent track record).  Nobody is going to want a 150 lb humanoid robot in their house designed by that company.

2.  There is a successful cyber attack that leads folks to question overreliance on automation (Every Cyber Attack Facing America).  AI could help bring this zero day hack to fruition and America is making a lot of enemies.  If that happens, see point 1 quite likely scaled up.

[Telecaster]

The robot taxi shouldn't cost _more_ than a current model 3, it doesn't have any extra sensors or equipment. So that should be less than $35,000/car (that's sale price, some of that is profit, most is cost of goods), while the hardware for each Waymo is estimated to be a couple of hundred thousand dollars. An even modestly less capable automated ride hailing service -- assuming it is "good enough", waiting to see if this is true for Tesla's solution or not -- might be able to undercut Waymo on price or improve on quality of service by having a higher density of cars available if each cost costs 1/7th as much.

TL;DR: consider quality per unit cost, not just absolute quality

FWIW, I just stumbled on this Bloomberg article that estimates Waymo's sensor suite cost at $9300.   

https://www.bloomberg.com/features/2025-tesla-full-self-driving-crash/

I'm not talking software. I'm talking about fit and finish.  Panel gaps, trim falling off, that sort of thing.  Magna is not well-known outside the industry,  but is the Foxconn of cars.  And, Foxconn itself wants to get into building cars.  The maximizing profit, nimble answer would be to outsource that kind of work.

Also FWIW, I believe Magna builds the I-Pace for Waymo.   

[maizefolk]

The robot taxi shouldn't cost _more_ than a current model 3, it doesn't have any extra sensors or equipment. So that should be less than $35,000/car (that's sale price, some of that is profit, most is cost of goods), while the hardware for each Waymo is estimated to be a couple of hundred thousand dollars. An even modestly less capable automated ride hailing service -- assuming it is "good enough", waiting to see if this is true for Tesla's solution or not -- might be able to undercut Waymo on price or improve on quality of service by having a higher density of cars available if each cost costs 1/7th as much.

TL;DR: consider quality per unit cost, not just absolute quality

FWIW, I just stumbled on this Bloomberg article that estimates Waymo's sensor suite cost at $9300.   

https://www.bloomberg.com/features/2025-tesla-full-self-driving-crash/

The estimate I was going off of was a quote from Dmitri Dolgov, Waymo’s co-chief executive, that the cost of all the equipment they have to add to a Jaguar I-Pace to turn it into a Waymo taxi is on the order $100,000 (that's not only sensors but electronics, wiring, computers, etc).* The cost of the base car itself is the order of $75,000. Plus the labor of retrofitting which is high right now but would likely come down at scale, or if/when google launches a car designed from the ground up to serve as a robotaxi.

*The quote was in a new york times article last September but it traces back to a podcast from February, so costs have doubtless come down somewhat in the last 16 months. https://www.nytimes.com/2024/09/04/technology/waymo-expansion-alphabet.html

[waltworks]

150 pounds seems really...optimistic for something that isn't tethered to a power source (unless... Tesla coils?!?) and has to walk around and move things at least as well as a normal human and have a pretty monster processor (and the battery to power it) and such. But who knows.

Musk's tendency for... exaggerating makes me believe the robots are a decade out, if they're a real product at all. So to some extent I'm not sure it matters.

-W

[ColoradoTribe]

Even if you can make a humanoid robot that can bag groceries, it's going to cost a heck of a lot more than a teenager, though, right? I build and maintain machines that are many orders of magnitude simpler (bicycles) and they still require quite a bit of work to keep running well. I can't even imagine the maintenance challenges involved with, say, a robot hand capable of handling a ripe tomato or something. Let alone the rest of the robot.

From an engineering standpoint, the likelihood seems really high that a hyper-complex device interacting with humans IRL is going to get damaged a lot. And in an industrial setting that's more controlled, there's not much reason for the humanoid form as others have already pointed out.

It's probably irrelevant regardless for quite a while. There's no evidence that Tesla has anything approaching a robot that can bag groceries, teleoperated drink-service notwithstanding.

Walt, I would suggest you look up videos of the latest versions of Optimus performing household taks or the one of Optimus dancing. If they are not already at the level that can pack a grocery bag without damaging any of the produce, they are very close.

As for the rest of your post, it's a pretty simple napkin math problem. Let's use stocking grocery shelves. That's a job that can and often is done 24 hours a day at a grocery store using shift work. Let's say the average shelf stocker makes $15/hr. That means to pay a human labor to stock shelves for an entire year you are looking at  $131,400 in labor. That's before we consider costs for sick leave, training, uniforms, an HR department, workers comp insurance, health insurance, paid holidays, etc. Alternatively, one Optimus robot will cost around $30k to start. Let's be generous and assume it need $10k/year in maintenance and uses $10k in electricity. That means a grocery store would recoup the cost of the robot in under six months and profit over $100k a year. How many grocery stores in the country. How many stockers per store? That's just one example. Labor is about to get very cheap.

[ColoradoTribe]

Manufacturing volume tend to go along with more optimized/lower maintenance designs. I've worked with research instruments that are essentially had build by skilled techs, limited run goods (say a thermocycler, or a precision four row air planter), and super widely produced items (cars, laptop computers, phones). Bigger production runs both means it makes sense to spend more time identifying and reducing potential failure/maintenance issues and that, with large number of devices out in the world, over time, you have more chances to identify what the actual failure and maintenance issues are that need to be addressed.

In the hypothetical world were any company had a low-ish cost robot capable of bagging groceries AND they were able to sell hundreds of thousands of those robots, I suspect that over time the maintenance requirements would look more like that of a car* than of a piece of research equipment or industrial automation. I certainly hope it'd look more like a car than like a cell phone/laptop where if anything serious breaks many people just throw it away and replace it.

*Regular periodic maintenance, occasional expensive repairs at a local shop with specialized mechanics when something breaks/wears out or when it experienced the robotic equivalent of being rear-ended or sideswiped.

You're getting into very important points. The thing that seems to sell cars is looks that give an impression of coolness, sexiness, strength, etc. Ask the people signing up for 7 year leases on pretty trucks. They thought it looked cool. They liked the facade redesign. They thought it would get them respect.

Will something similar apply to $30,000 AI-driven robot-people marketed to the masses? Or will the "general" robot market be dominated by the question "is it profitable to use?" the way the specialist robot market already is?

I suppose the case for Optimus rests on the assumption that people will buy humanoid robots for the same reasons they buy brand new cars. A specialist robot like the arms that weld car bodies together or the pucks that vacuums floors will always be cheaper to do any specific function, but the AI integration would supposedly make it possible for a general robot to go from welding to bagging groceries to helping nursing home patients to vacuuming the cat hair.

Then we must assume that all those joints, servos, motors, cameras, sensors, processors, batteries, hydraulic systems, etc. are durable and low-maintenance enough so that the robot can earn back its cost over a period of perhaps decades. And the long-term energy bill needs to be evaluated too. What are we looking at? 500 watts just to stand there? These machines would need to be produced to an extraordinary quality level to financially justify themselves... unless they are bought as vanity products, in which case they can be as disposable as a modern Audi SUV or set of chrome pretty truck rims.

Yet, I'm also missing their surveillance value. The reason you can buy a 65" 4K television for under $400 is because the TV is spying on you, sending advertisers information about what you're watching so that ads can be targeted to your specific demographics and interests. Perhaps your robot will be similarly discounted, but will steer you toward purchases, advertisements, fake news sources, or subscription upgrades. Perhaps we'll accept ad-subsidized robots in our lives the same way we accept it from our TVs and phones?

This makes little sense. A car is a depreciating asset. A humanoid robot that provides general labor at a fraction of the cost of a human laborer is an investment, one that pays for itself in months before turning into a money printing machine. In my above example, Safeway doesn't care about "sexy", they will care that the robot can stock a grocery shelf for a fraction of the cost.

Per Grok, "BLS data for "production" occupations (e.g., manufacturing) shows total compensation at $37.16 per hour, with wages at $25.89 and benefits at $11.27.
Assuming an 40-hour workweek and 52 weeks per year (2,080 hours), annual costs for a manual labor employee are approximately $77,293, with $53,851 in wages and $23,442 in benefits."

However, in my example, one Optimus can replace multiple human shifts that are fulfilling the same job function, since robots don't get union breaks, don't go to the bathroom, don't smoke, don't get injured and don't "sleep" longer than it takes to recharge. So, for a robot filling a shift position on a line, we can double that $77k to $144k and still have plenty of  time for recharging, maintenance and repairs.

Optimus has an anticipated price of ~$30k. Maintenance and charging for a year is not going to more than double that figure. So, assuming a generous $50K in year one and the robot pays for itself in under six months and nets around $120k in years 2 through 10. If folks want to assume a 5 year life expectancy that's fine, basic math is unchanged. The disruption is coming.

[Telecaster]

So while I don't know which way the economics would actually work out in this particular scenario (and I also don't know if the generalizable humanoid robot is going to be possible to create and/or manufacture at scale), I think it is at least conceptually possible for a more complex general purpose robot that can do 500 different tasks and is manufactured on a massive scale might end up being cheaper for end users than designing 500 different simpler specialized robots, each with much more limited production runs where the costs of both R&D and capital investments like building assembly lines need to be spread out over fewer units.

There is a lot of merit to this line of thinking.  One of the great tech business innovations was Texas Instruments decision to mass produce chips.  Previously, they would produce enough chips to meet a single customer's order, so the cost had to be amortized over a small number of chips.   But they decided to try just pumping them out, and as the price went way down, people starting finding new ways to use the chips.   So they induced their own market demand.

Certainly if general purpose robots were cheap enough people would start finding things for them to do.   But we're aways off, I think.    Humanoid robots have about a 1-2 hour battery life.    That is far too short for general industrial use.   That means you need to plug them in, which means you don't need them to be mobile (or not very anyway).   Substituting wheels for legs increases battery life and reduces costs, but same basic problem.   If you have to tether them to a work station, then you probably don't need a general purpose robot.   Purpose built or a GP robot modified for the task is probably more suitable.   

And I believe at least in the early years there will still be the need for customization.   A lot of welding is still done by humans.  A welding robot presumably would need the welding optics package, extra good fine motor control, and maybe an extra set of arms to manipulate the work piece.  A fabric working robot might need sensors in the fingers to detect squishy things, but a dishwashing robot might need the firm non-slip grip option.    A window washing robot might need upgraded optics processing so it can detect glass and tell how dirty a window is.  That sort of thing.   Maybe one day they would all merge, but I think that day is down the road still.

I was thinking a humanoid robot would be good at doing wildlife surveys.   It could do a perfect grid pattern and accurately identify and document all of the flora and fauna in an area.   Same with timber surveys.   

[ColoradoTribe]

The robot taxi shouldn't cost _more_ than a current model 3, it doesn't have any extra sensors or equipment. So that should be less than $35,000/car (that's sale price, some of that is profit, most is cost of goods), while the hardware for each Waymo is estimated to be a couple of hundred thousand dollars. An even modestly less capable automated ride hailing service -- assuming it is "good enough", waiting to see if this is true for Tesla's solution or not -- might be able to undercut Waymo on price or improve on quality of service by having a higher density of cars available if each cost costs 1/7th as much.

TL;DR: consider quality per unit cost, not just absolute quality

FWIW, I just stumbled on this Bloomberg article that estimates Waymo's sensor suite cost at $9300.   

https://www.bloomberg.com/features/2025-tesla-full-self-driving-crash/

The estimate I was going off of was a quote from Dmitri Dolgov, Waymo’s co-chief executive, that the cost of all the equipment they have to add to a Jaguar I-Pace to turn it into a Waymo taxi is on the order $100,000 (that's not only sensors but electronics, wiring, computers, etc).* The cost of the base car itself is the order of $75,000. Plus the labor of retrofitting which is high right now but would likely come down at scale, or if/when google launches a car designed from the ground up to serve as a robotaxi.

*The quote was in a new york times article last September but it traces back to a podcast from February, so costs have doubtless come down somewhat in the last 16 months. https://www.nytimes.com/2024/09/04/technology/waymo-expansion-alphabet.html

I did a detailed post about all the components a model 3 has that a robo-taxi doesn't need or have. The COGS for a base Model 3 is around $35k and Tesla could get that down to $25k for the robotaxi without breaking a sweat. Just off the top of my head, no expensive glass ceiling, no turn stalk, no steering wheel, no pedals, two doors instead of four, two seats instead of five, no rear view or side view mirrors, cheaper wheels, cheaper sound system, and less wiring. Perhaps most importantly, a smaller battery. A robotaxi doesn't need 350 miles of range.

Maize is right about the cost of a Waymo being around $175K. Think about this. Tesla robotaxis will start rolling off a limited production line this year and will have the software, AI brain and hardware needed to function as an AEV the minute it comes off the line and at a cost of around $20k-$25K per vehicle.

Alternatively, Waymo has to buy an existing EV (currently I-Pace, but switching to Ioniq), then Waymo has to buy the sensor suite of radar, lidar, etc., then both the cars and components are shipped to yet another company to have the hardware components installed, and lastly Waymo installs their software. Every step and vendor adds to the cost and complexity of the end product. Waymo presumably can't repair a broken sensor or a broken down EV on their own either.

The second Tesla starts safely providing rides to the paying public (sometime this month in Austin), Waymo is a dead-man walking. A competitor that costs 7X up front to make, that can't operate on highways, is not vertically integrated, and has to map and geofence each service area independently is not a competitor.

[ColoradoTribe]

So while I don't know which way the economics would actually work out in this particular scenario (and I also don't know if the generalizable humanoid robot is going to be possible to create and/or manufacture at scale), I think it is at least conceptually possible for a more complex general purpose robot that can do 500 different tasks and is manufactured on a massive scale might end up being cheaper for end users than designing 500 different simpler specialized robots, each with much more limited production runs where the costs of both R&D and capital investments like building assembly lines need to be spread out over fewer units.

There is a lot of merit to this line of thinking.  One of the great tech business innovations was Texas Instruments decision to mass produce chips.  Previously, they would produce enough chips to meet a single customer's order, so the cost had to be amortized over a small number of chips.   But they decided to try just pumping them out, and as the price went way down, people starting finding new ways to use the chips.   So they induced their own market demand.

Certainly if general purpose robots were cheap enough people would start finding things for them to do.   But we're aways off, I think.    Humanoid robots have about a 1-2 hour battery life.    That is far too short for general industrial use.   That means you need to plug them in, which means you don't need them to be mobile (or not very anyway).   Substituting wheels for legs increases battery life and reduces costs, but same basic problem.   If you have to tether them to a work station, then you probably don't need a general purpose robot.   Purpose built or a GP robot modified for the task is probably more suitable.   

And I believe at least in the early years there will still be the need for customization.   A lot of welding is still done by humans.  A welding robot presumably would need the welding optics package, extra good fine motor control, and maybe an extra set of arms to manipulate the work piece.  A fabric working robot might need sensors in the fingers to detect squishy things, but a dishwashing robot might need the firm non-slip grip option.    A window washing robot might need upgraded optics processing so it can detect glass and tell how dirty a window is.  That sort of thing.   Maybe one day they would all merge, but I think that day is down the road still.

I was thinking a humanoid robot would be good at doing wildlife surveys.   It could do a perfect grid pattern and accurately identify and document all of the flora and fauna in an area.   Same with timber surveys.

Second generation Optimus robot has a 3 kw battery pack (4680 cells) that provide between 8-10 hours of operating time. So, long enough to complete a standard human work shift but without the need for breaks.

[maizefolk]

Okay Telecaster, I think we're on the same page then.

If it were possible to make a robot that could (without the need for customization) do a large range of tasks, it might be the case that economics of scale would bring down the cost of that robot enough that it would make more sense to use a more complex general purpose robot for tasks where right now we use simpler single purpose robots or actual human labor.

I'm not trying to convince anyone it is possible with current tech to build such a robot at a low enough cost that the positive feedback loop of "economically viable for more tasks" -> "more demand/more scale" -> "lower cost" -> "economically viable for more tasks" kicks in, in part because I'm not confident of that myself.

[neo von retorch]

https://arstechnica.com/ai/2025/05/hugging-face-hopes-to-bring-a-humanoid-robot-to-market-for-just-3000/

Good news - they can be as cheap as $3000.
Bad news - they can be as cheap as $3000... without having to go to Tesla to buy one!

While past investment of Tesla was very profitable, and as a meme stock, it's impossible to predict the price going forward,

I find that the "investment analysis" that is based on the rosiest pictures put forth by Musk while ignoring all possible competitors existing is where things seem to fall short. Ignoring the fluffed up claims, and looking at historical reality, where's the real world moat?

[MustacheAndaHalf]

I think Optimus will take a lot longer to sell than Musk thinks.  There will be rich, early adopters who just want a robot... but then what?  Liberals tend to score higher on "openness to experience", which is what you need to try out an AI robot.  But Musk has aligned himself closely with Trump, alienating himself from that customer base.

There's a book about innovative products called "Crossing the Chasm", which talks about the wide gap between early adopters and the mainstream customer.  I think Optimus could easily fall into that chasm.

There’s a reason most robots sold today look nothing like a human body. That reason is that we have plenty of human bodies - billions of them, and if you hire them to do a task you can just pay them by the hour rather than making a large upfront investment in purchasing a thing (or various lease agreements in which this cost is factored in).

The bestselling robots today are things like 12 foot arms that make spot welds, forklift-like robots, robots that fit inside pipes, and airborne drones. In other words, things with a radically different form factor than the human body.

Honda’s Asimo humanoid robot from a decade ago was arguably more advanced than Tesla’s remote controlled Optimus, as it could autonomously respond to various cues. Still, it failed to find a market. This suggests to me that Optimus is more about generating sci fi excitement for the stock price than meeting any particular market need. It reminds me of The Boring Company.

Agreed - robots are already present in manufacturing and other places where a humanoid robot wouldn't be efficient.  Tesla's general robot will struggle to compete with specialized industrial robots.

According to this Tesla stock analyst, Optimus robots are already used by Tesla in its manufacturing.  Reading between the lines, it sounds like this analyst thinks Tesla could build up AI capability through experience with Optimus - perhaps helping some other products.  I don't see any discussion of humanoid robot sales.
https://finance.yahoo.com/news/analyst-revisits-tesla-stock-price-114110040.html

Did you not read the posts up thread. There are currently in the world 300 million humans working in factories that do some form of manufacturing. These are human manufacturing jobs where there is not a non-humanoid robot alternative. It is pretty obvious that the bulk of these 300 million human factory workers could be replaced by humanoid robots and the robot replacement would pay for itself in short order. If you have a line job being done by humans in three 8-hours shifts per day. Even if the worker is only earning $10/hr, that's a labor cost of $87,600/year to staff that job with humans. Optimus will cost around $30k to start and go down in cost from there with economies of scale. There are literally millions of factory workers whose jobs can be done more safely and cheaply by a humanoid robot/AI.

China has the most factory workers, where they are paid an average of $8,200 USD per year, or less than 1/10th of your guess.  Your guess is wildly off - you should have questioned why you estimated factory workers to earn higher wages than the average American.  The robots cannot do those jobs more cheaply.
https://www.erieri.com/salary/job/factory-worker/china

Door man

You accused me of not reading prior posts, while yourself ignoring prior posts:

But why do doormen exist at all?  It's not really to open doors.  They're there partly as a 'look how rich we are, we can hire a dude just to stand there and look pretty' flex, partly do to random general tasks - call a taxi, carry boxes, act as security (checking if people are tailgating when residents enter), sign for/accept packages for guests, answer information queries, general cleaning if the weather outside is poor and the entrance-way gets wet and needs to be mopped.  I don't think Optimus would fill this use case very well at all.

[reeshau]

Just a couple random thoughts -

1.  Tesla kills someone during the limited Austin rollout of FSD (not far fetched given their relatively recent track record).  Nobody is going to want a 150 lb humanoid robot in their house designed by that company.

2.  There is a successful cyber attack that leads folks to question overreliance on automation (Every Cyber Attack Facing America).  AI could help bring this zero day hack to fruition and America is making a lot of enemies.  If that happens, see point 1 quite likely scaled up.

I'm just waiting to hear about Tesla's service on the first day there is a "gully washer" rainstorm in Austin.  It will be a significant disincentive to have a rideshare service that can't operate in the heavy rain, or that has to interrupt rides in progress.

It can be 99% perfect.  But if that 1% is fatal, to the service or the car, it will be a differentiator or could kill the service.  And yes, the brand identity will transfer to the other products.  Something like that might not be immediately fatal, but it certainly gives openings to competitors.

[ChpBstrd]

Michael Cohen thinks Trump will target Musk by attacking Tesla and clawing back all the subsidies the business was built on.
https://www.rawstory.com/michael-cohen-trump-2672247489/

[dividendman]

Michael Cohen thinks Trump will target Musk by attacking Tesla and clawing back all the subsidies the business was built on.
https://www.rawstory.com/michael-cohen-trump-2672247489/

Let's go Trump lol, my puts need it.

[Paper Chaser]

The grocery bagging robot example makes me chuckle. Why would low-margin grocery stores add that expense when customers bag their own for free right now? Every big box store that I enter has increased the number of self-checkouts in the last 5 years, while reducing the number of costly employees.
My local Walmart was remodeled last year, and went from having ~30 traditional checkout lanes + ~12 self checkout registers to having a corral of ~30 self checkout lanes and 0 traditional registers. They have 2 employees at the entry point directing customers to open stations and 2 at the exit watching for shoplifting (probably with some help from cameras/AI).
Convenience stores are offering cashierless transactions, where you just place your items on a table and they're automatically scanned. Amazon has used cashierless stores for awhile.

The same thing is happening in manufacturing, where smarter systems can improve safety, reduce human effort, improve quality, etc.

Low-skill, menial jobs are definitely at risk, but I don't think that risk is humanoid robots. It's just low-level automation, better systems, and expecting customers or a smaller number of employees to do more of the labor.

Even in dangerous situations like the military, we see trends of reducing human involvement. Manned aircraft, boats, etc are being replaced by remote controlled drones. We're not sending humanoid robots into warzones to fight each other. We're flying super cheap drones around remotely.

[reeshau]

I think a good example of grocery store thinking is seeing the floor sweeping machines strapped with automation equipment to be autonomous (a commercial-scale Roomba) and to do inventory checks with a side scanner.

It wasn't even a purchase, it was a retrofit.   They are, and need to be, THAT cheap.

[ColoradoTribe]

The grocery bagging robot example makes me chuckle. Why would low-margin grocery stores add that expense when customers bag their own for free right now? Every big box store that I enter has increased the number of self-checkouts in the last 5 years, while reducing the number of costly employees.
My local Walmart was remodeled last year, and went from having ~30 traditional checkout lanes + ~12 self checkout registers to having a corral of ~30 self checkout lanes and 0 traditional registers. They have 2 employees at the entry point directing customers to open stations and 2 at the exit watching for shoplifting (probably with some help from cameras/AI).
Convenience stores are offering cashierless transactions, where you just place your items on a table and they're automatically scanned. Amazon has used cashierless stores for awhile.

The same thing is happening in manufacturing, where smarter systems can improve safety, reduce human effort, improve quality, etc.

Low-skill, menial jobs are definitely at risk, but I don't think that risk is humanoid robots. It's just low-level automation, better systems, and expecting customers or a smaller number of employees to do more of the labor.

Even in dangerous situations like the military, we see trends of reducing human involvement. Manned aircraft, boats, etc are being replaced by remote controlled drones. We're not sending humanoid robots into warzones to fight each other. We're flying super cheap drones around remotely.

Grocery bagger was an example of a job that may come back if labor costs drop to $2/hr or thereabouts based on the lifetime TCO of a humanoid robot divided by the number of labor hours it provides.I could definitely see higher end chains competing for customers in this way. But, if you din't like that example, see my example using the grocery shelf stocker. There's not a single specialized, non-humanoid robot that could handle that job.  A humanoid robot with AI could perform this task though and at a fraction of the cost of human labor.

[ColoradoTribe]

Just a couple random thoughts -

1.  Tesla kills someone during the limited Austin rollout of FSD (not far fetched given their relatively recent track record).  Nobody is going to want a 150 lb humanoid robot in their house designed by that company.

2.  There is a successful cyber attack that leads folks to question overreliance on automation (Every Cyber Attack Facing America).  AI could help bring this zero day hack to fruition and America is making a lot of enemies.  If that happens, see point 1 quite likely scaled up.

I'm just waiting to hear about Tesla's service on the first day there is a "gully washer" rainstorm in Austin.  It will be a significant disincentive to have a rideshare service that can't operate in the heavy rain, or that has to interrupt rides in progress.

It can be 99% perfect.  But if that 1% is fatal, to the service or the car, it will be a differentiator or could kill the service.  And yes, the brand identity will transfer to the other products.  Something like that might not be immediately fatal, but it certainly gives openings to competitors.

Have you ever ridden in a Tesla using FSD v13 during a rainstorm? I'm curious. Have you ever be driven around on v13 period?

I have and it was a non-issue. I rode through heavy rain and the car performed flawlessly. Is there a point where even v13 can't operate in a torrential downpour? Of course, but it's at roughly the same point a human driver should put on the hazards and pull over to wait it out.

[ColoradoTribe]

I think a good example of grocery store thinking is seeing the floor sweeping machines strapped with automation equipment to be autonomous (a commercial-scale Roomba) and to do inventory checks with a side scanner.

It wasn't even a purchase, it was a retrofit.   They are, and need to be, THAT cheap.

Can the retrofitted Roomba pick-up a can of baked beans and put it in the proper spot on the top shelf, while also differentiating between honey bourbon and bacon honey baked beans?Seems like you guys are forgetting the AI part of the robot. The AI part opens up a whole other set of jobs and tasks that a specialized robot or a robot without AI can't perform, because the Ai robot can quickly be taught new or very specialized tasks.