Remember back in 2021 when Musk showed off his Johnny 5-esque Tesla Robot at their AI day? As with all of his live events, Musk’s PR slant was in full swing, claiming that the bot and its FSD-based AI computer could do all manner of manual labour, from assembly line work to even picking up your groceries. He even claimed it would go into production in 2022… Since then, he has claimed that the market demand for his robot could reach 20 billion units! Yet, in Tesla’s most recent demo of their Gen 2 bot, dubbed “Optimus”, it seems no more capable than any other humanoid robot, such as those from Boston Dynamic or the world-famous ASIMO. Yet, despite Musk’s rich history of vastly over-promising and severely underdelivering, some serious movers, like Morgan Stanley and Goldman Sachs, are predicting that Musk’s robot will spark an utter game-changing revolution! Sadly, this doesn’t seem to be the case at all. But if you dig a little deeper, you will find the disillusioning reason for not only the creation of Optimus, but also why Morgan Stanley and Goldman Sachs are doggedly backing it.
Make no mistake, Optimus is an impressive robot. During the recent demo, it displayed its incredibly useful strength and remarkable dexterity. Moreover, the fact that it uses the same computer and visual AI that drives Tesla’s Autopilot means that Optimus can readily understand, navigate and interact with its environment surprisingly well and should learn tasks relatively easily. So yes, while it may still have that classic robot “I’ve just crapped my pants” walk, it is still one of the more advanced robots out there.
As such, the investment bank Morgan Stanley predicts that Optimus can disrupt 30% of the global labour market. That equates to a mind-blowing $30 trillion opportunity. Goldman Sachs, another investment bank, is slightly more restrained, claiming that Optimus could be a $150 billion business in the next fifteen years!
** Quick interruption, if you want more from me, or to interact with me, go follow me on Bluesky, X, or Substack**
As Musk claims Optimus will cost “much less than $20,000” per unit, these figures strongly suggest that these banks expect the robots to disrupt job markets far beyond just basic manual labour. Factory jobs and even skilled but “constrained” jobs like mechanics are all at risk. So, as always, cue terrified workers fearing for their jobs’ security.
But, the reality is far from this dystopian utilitarian ultra-capitalist dream of the future.
Take the computer and AI at the heart of Optimus. The computer is identical to the ones inside a Tesla, and the AI is directly based on Tesla’s self-driving AI. But Tesla has yet to actually make a proper, reliable, legal, fully self-driving car. In fact, Tesla and its self-driving system are being investigated by the DoJ and are even under an NHTSA probe because of a series of serious crashes, some fatal, linked to the use of Tesla autopilot. So, right now, they are still quite a way off full autonomy.
Now, driving a car is a far more restrained task than even the most simple unskilled job. There are fewer variables, more apparent rules to follow, and far easier to execute tasks than, say, if Optimus was used as a construction site labourer or a cleaner. Particularly if Optimus has to go to different locations and learn the area from scratch each time.
So, If Tesla can’t produce a proper self-driving car, how can their self-driving computer and AI system be used to make a functional general-purpose humanoid robot?
This is made even more worrying when you look at recent Tesla whistleblower Lukasz Krupski’s claims and leaked data. He has shown that Tesla’s Autopilot has an issue with anomalous events like “phantom braking”, which Tesla is actively not looking into. In fact, during a recent test drive, I experienced such an event. This lack of reliability within the AI can be catastrophic when used to drive a car, but it can be equally destructive when used in a robot. What if it is carrying a heavy load around a building site and drops it on someone due to an anomalous event?
Now, unskilled workers suffer a vast amount of such accidents each year, and there is an argument that using Optimus in these situations can mitigate human injury or deaths. But that has yet to be proven, particularly if Optimus works alongside humans. But when it is just humans, who is liable is relatively obvious. This isn’t the case with Optimus in the mix. Is it the fault of Tesla? What about the guy who trained the AI to do the task? Or is it the owner of Optimus? Or, is there no entity liable for damages, both fiscal and health-wise? What if Optimus could have stopped a human-caused injury, death or loss but didn’t? Can whoever is liable be found at fault here?
There is also an issue of qualification. Let’s say Optimus can pass a skilled job qualification, like a bricklayer. Does the entire fleet of robots get the qualification? Or just that single robot? The AI is constantly changing through retraining, but these changes are not always in a better direction, meaning Optimus could, theoretically, get worse at tasks, at least briefly. Does that mean that when the AI gets its regular updates, is it a different program with different abilities? And therefore, does it need to retake its qualification? This could get tiresome if these updates come relatively regularly.
There is also a considerable problem with efficiency. You see, humanoid general-purpose robots simply aren’t efficient. It’s more efficient to use a Roomba than to buy a Tesla Bot and a Dyson to hoover the house for you. Using articulated arms in a factory is more efficient than using an Optimus to assemble products. Most unskilled labour can be serviced by specialised robots that are more reliable, cheaper and less energy-intensive than a general-purpose humanoid robot. This is why many robotics professionals don’t bother with humanoid robots.
The same is true of AI. Currently, in very constrained parameters and applications, it can be brilliantly powerful, such as the AIs that calculate the structure of proteins, discover new antibiotics, or even AI-aided design. But, as of yet, AI struggles with general applications. Take ChatGPT and other large language models. They can be very passable if you want them to talk in broad terms. But get them to talk about specifics or write articulate articles, and they make facts up (even ChatGPT 4), get references wrong or sometimes just don’t make all that much sense. So, how can AI power a general-purpose robot in its current guise when it is far from general-purpose?
It’s not just me saying this. Dennis Hong, a professor of mechanical & aerospace engineering (otherwise known as a robotics expert) at UCLA, said of the recent Optimus demo, “I disagree with is their decision to mass produce the model shown at the event. I’m sure there are many good reasons behind it that I am not aware of, but I do not believe that is the one that can be used in a real-world setting in any meaningful capacity.” He continued, “Though it may be used in a structured environment in a limited way (very simple tasks in a very carefully controlled factory maybe?), this particular model still lacks some of the fundamental technology for locomotion & manipulation needed for use in a general situation.”
In other words, as it currently stands, Optimus is a brilliant party piece, but as an actual robot, it is far from viable or efficient.
So why are Musk, Bloomberg and Goldman Sachs so bullish about Optimus?
The answer, as I have pointed out many times in the past, is easy. They all have skin in the game.
Musk has stated numerous times that the reason Tesla is valued so highly is because of its self-driving AI. In fact, many serious investors and stock analysis groups agree with him. You see, if (and that is a big if) Tesla can master self-driving software, then every Tesla on the road can become fully autonomous. This could drive Tesla’s profits through the roof, from operating their own robotaxis to current Tesla owners paying to upgrade to fully autonomous driving to new customers flocking to Tesla.
So basically, Tesla’s stock price is so high because investors speculate that in the future, Tesla could be worth far more!
But, as of yet, Tesla is far from reaching full autonomy. Mercedes and Audi have self-driving cars that are already legally operating at higher levels of autonomy than Tesla. In fact, there is a lot of evidence that since Tesla switched to vision-only Autopilot in 2021 (which only uses cameras, no other sensors), their self-driving AI has actually gone backwards and lost capability and safety. So, if you were to value Tesla on its current AI status (which we will do in a minute) and development prospects, it would be far lower than its current market cap.
This is why Musk started the Tesla Bot project and is talking about selling billions of units. There is basically zero evidence that he, or this technology, can reach this goal. Instead, Optimus is meant to be a platform to show off Tesla’s self-driving computer and Tesla’s AI. It might not be able to function as a real general-purpose robot (or AI), but it can help generate hype around Tesla’s AI and keep it in the media. In other words, it’s a party trick to artificially (in my opinion) inflate Tesla’s stock value. This is made all the more questionable when you consider that Tesla’s 2021 switch to vision seems to have slowed down Tesla’s self-driving development quite significantly.
But how much is this hype around Tesla’s AI driving up the stock price? Well, thanks to Toyota, we can see.
You see, Toyota also has AI self-driving technology at the same level as Tesla, and it is also heavily developing fully autonomous vehicles. But it isn’t just AI that Toyota has going for it; it also has game-changing EV technology coming very soon in the form of a 745-mile capable solid-state battery. It even has its own version of Tesla’s gigacasting (which makes car production even cheaper). Toyota has an almost identical profit margin to Tesla of 17% (Tesla is at 17.89%). However, this was a 5-year low for Toyota, and Toyota has always sold way more vehicles than Tesla each year.
So, what if we valued Tesla the same way that Toyota is? They are very similar companies, with very similar future prospects, just Tesla has a higher perceived (but not current) value in its technology.
Well, Toyota made a whopping $57.167 billion in gross profit during the 2022–2023 fiscal year. At the end of this fiscal period, Toyota’s market cap (value of all its stock) was $247.03 billion. This gives Toyota a profit/market cap ratio of 40.76. During the same period, Tesla made only $18.99 billion in gross profit, meaning Tesla made 67% less than Toyota. Despite this, at the end of this fiscal period, Tesla had a market cap of $774.41 billion!
If we apply Toyota’s profit/market cap ratio to Tesla, we can estimate what Tesla is worth solely based on its current profit, sales and AI, as they are very similar to Toyota. If you do the maths, Tesla’s market cap comes out at only $82 billion. That’s nearly 90% less than what it is worth today!
So, you can see that for Musk, keeping the perceived value of Tesla’s AI is vital! If this hype train slows down, his business could lose hundreds of billions in value.
But, this is particularly important to Musk personally, too, as he uses his 12.5%, $90 billion stake in Tesla as collateral for multi-billion dollar loans. If Tesla was valued the same way as Toyota, his shares would only be worth $10.25 billion, meaning his money printing machine would dry up rapidly.
So, that’s Musk; what about Morgan Stanley and Goldman Sachs?
Well, both of these investment banks are heavily invested in Tesla. Morgan Stanley holds, as of 2023–09–30, 30,160,747 shares in Tesla, with a current value of roughly $7.5 billion. They have significantly increased their stakeholders in Tesla throughout 2023. Goldman Sachs is in a similar boat. As of 2023–11–14, they owned 15,605,198 shares in Tesla, with a current value of roughly $3.9 billion. Like Morgan Stanley, they have dramatically increased their stake in Tesla throughout 2023. These banks seem to have taken advantage of some of the dips in value over the year to boost their stake on the cheap.
But this means that if Tesla’s AI hype slows down, Morgan Stanley and Goldman Sachs could stand to lose literally billions of dollars!
As such, it really shouldn’t be a surprise that they are egging on Tesla’s AI party piece and making baseless claims about how impactful and profitable it could be. This marketing spin and PR on behalf of Musk lines their own pockets.
So, in summary, yes, Optimus is an impressive robot, and it could have some meaningful impact on the world. However, its prospects are incredibly limited and far from what Musk is promising due to the limitations of robotics and AI. Yet, Musk and all of his major stakeholders are using Optimus to drive AI hype around Tesla to keep their investment value inflated. Which is a huge shame. If Tesla’s AI is as powerful as Musk claims, then why isn’t he using it in applications that we know AI can be incredibly productive? Such as CAD, drug development, environmental monitoring/prediction, Healthcare such as cancer diagnosis, supply chain optimisation and even energy grid design and management. These could demonstrate just how powerful their AI is whilst actually having a viable commercial application, many of which even fit into Tesla’s current business model. In other words, it could make Tesla’s stock genuinely worth more. I think it is telling that Musk and Tesla aren’t taking this approach and instead are using a more smoke-and-mirrors PR stunt to increase their value.
Thanks for reading! Content like this doesn’t happen without your support. So, if you want to see more like this, don’t forget to Subscribe and follow me on BlueSky or X and help get the word out by hitting the share button below.
Sources: Forbes, Will Lockett, Yahoo Finance, Cleantecnica, Cleantecnica, Manchester University, Fintel, Fintel, CMC, Macrotrends, Macrotrends, Y Charts, Fin Box, NASDAQ, Electrek, Statista, Will Lockett, Techopedia, NY Times, Toyota, The Data Science