Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

[Mardak]

They tried to copy Tesla and made a big mess.

I'm not suggesting legacy auto needs to do something smart or clever. In fact, being a dumb manufacturer is probably more valuable to the other companies that are developing / investing in autonomy software.

Probably not possible because they have stockholders that expect profits and dividends.

Existing auto manufacturers provide value in being able to manufacture the many millions of vehicles needed to collect data for improving autonomy software as well as the actual robotaxis. Once it becomes clear to the stockholders that traditional business of selling cars to human drivers is going to lose to robotaxis, there could be a quick shift or bankruptcy.

Only risk are those from FANG who are able to raise enough capital if not more to go against Tesla.

For those who realize the value of robotaxis, say Apple or Alphabet/Waymo, they can and have been putting in lots of money in potentially the "wrong" approach, e.g., expensive LIDAR and pre-mapping, so once they see proof of Tesla's vision-only solution deployed, shifting that money to directly pay the upfront costs of installing cameras and compute on some "dumb" auto manufacturer's vehicles would be relatively cheap and could result in actually achieving their own cost competitive robotaxis.

I still don't understand how anyone thinks that there will be fast followers in autonomy.

It's the same problem Tesla has always had -- relatively limited production. Elon Musk realizes this when he says robotaxi introduction is slowed down by manufacturing constraints. If Tesla can avoid that by rapidly expanding their own production as well as leveraging production of other auto manufacturers (who need to realize sharing profits with Tesla is better than gambling on other robotaxi approaches or bankruptcy), that could indeed prevent much of the potential fast followers from eating into robotaxi marketshare.

 

[LiveLong&Profit]

Something I haven't seen talked about nearly enough, and that I think is being badly missed in the autonomous car product / revenue discussion is that at the end of the day its still transportation. And transportation is a commodity business with razor thin margins. I absolutely see a short term (3-10 years?) "golden era" for owners of transportation services where the competition will be with the current world order - personally owned vehicles; UBER/Lyft as currently constituted.

But as the number of robotaxis increases the day rapidly arrives where the current world order is no longer the competition - it becomes the new robotaxi product and service.


A thought experiment - if we grant that a robotaxi generates $500k of revenue after fuel, off of a $50k vehicle that needs another $50k in maintenance (simplifying somewhat to make the point easier), and Tesla takes 30% of that, then we've got a $350k return on $100k investment. I'll take 20 please.

BUT when we shift to the new world order where robotaxis will be competing with each other, who here is willing to lower that $350k return on $100k to say $300k? And if it were $300k, who is willing to lower it down to $250k?

This is from the trucking industry and 2018:

For the Driver

For the Driver from American Trucker serves independent owner-operators and growing trucking fleets operating fewer than five power units.

www.trucker.com

But it is at least directionally accurate and has the industry profit margins UP at 6% after previous years in the 2-4% range.

Two conclusions I draw:
1) These are the sorts of profit margins I see robotaxis evolving to. If I'm right this will be a lot of revenue and very little profit for Tesla. At most I hope Tesla would form a new division and then spin it out as a new business at these profit levels.
2) I see little to no way for Tesla to continue to take 30% off the top and leave owner / operators with 2-6% profit margins. If nothing else, some other company that gets a robotaxi into service could decide to take 25% off the top. Then Tesla lowers to 20%. And the race to the bottom is on until, my prediction, Tesla (or other company) is getting 2-6% profit margin and the owner / operator is earning 2-6%. I.e. - roughly an even split of the profits. Clearly with some businesses doing better and some worse.


This might take 20 years or 2 years in individual geographies as its deployed, but I see the outcome as inevitable. Its in the nature of transportation - technological advances and cost improvements flow through to the customers and very little sticks to people buying and deploying the technology because the business is just too big and the competition is too intense for anything else to happen.

For sure, competition could and would undercut Tesla. But when will they have vision-only robotaxis? 1 year? 2 years? 5 years?

I see a scenario, where Tesla is both first mover, and has the lowest cost (10 cent per mile) and pushes prices down year after year. Not only to destroy the price umbrella of the competition but because serving more and more customers align with the mission which is to electrify transportation (and power transportation by renewables and batteries).

In that case, Tesla does both an iPhone-scenario and an android-scenario - simultaneously.
Tech-leader, cost leader, brand leader, most robotaxies in action: How to compete against that?

And, as other have recounted, once you get below 70 cent per mile, then RT is cheaper than owning a car. So you expand the market by some wild scaling factor. Tesla can expand the market and just keep reducing prices. There is a long way from 70 cent down to 10 cent, and a lot of profit.

Eventually when you get close enough to 10 cent/mile you get to compete with public mass transit. And that unlocks a new large market.

 

[EVWatcher]

1. The Chinese people have no say in this. Xi will decide, given the policies he has undertaken so far I fear he may go for it.
2. China was militarily weak, they were humiliated in 1996 when they sabre rattled and the US sent a carrier between China and Tawain and they realised they couldn’t do anything. So they started a modernisation drive (which now ironically they could fund as a result of opening up the economy in the 1980s) Now that modernisation is paying off.

I don’t want to get to much into the possible reasoning behind a potential war other than to inform the shareholders on this board to factor this into account and to watch this closely because if a war does start it will probably be a surprise attack and straight away the stock will tank. Irrespective of what the US does in response.

The structure of chinas government does allow Xi to do whatever he wants. But I don’t think Xi or the Chinese government thinks war would be a productive thing to do especially given how fragile the economy is right now. As with all governments chinas governments control is dependent on stability. Xi and China are incredibly practical. Theres not enough to gain in going to war to take Taiwan considering what’s at stake.

 

[deshkart]

Something I haven't seen talked about nearly enough, and that I think is being badly missed in the autonomous car product / revenue discussion is that at the end of the day its still transportation. And transportation is a commodity business with razor thin margins. I absolutely see a short term (3-10 years?) "golden era" for owners of transportation services where the competition will be with the current world order - personally owned vehicles; UBER/Lyft as currently constituted.

But as the number of robotaxis increases the day rapidly arrives where the current world order is no longer the competition - it becomes the new robotaxi product and service.


A thought experiment - if we grant that a robotaxi generates $500k of revenue after fuel, off of a $50k vehicle that needs another $50k in maintenance (simplifying somewhat to make the point easier), and Tesla takes 30% of that, then we've got a $350k return on $100k investment. I'll take 20 please.

BUT when we shift to the new world order where robotaxis will be competing with each other, who here is willing to lower that $350k return on $100k to say $300k? And if it were $300k, who is willing to lower it down to $250k?

This is from the trucking industry and 2018:

For the Driver

For the Driver from American Trucker serves independent owner-operators and growing trucking fleets operating fewer than five power units.

www.trucker.com

But it is at least directionally accurate and has the industry profit margins UP at 6% after previous years in the 2-4% range.

Two conclusions I draw:
1) These are the sorts of profit margins I see robotaxis evolving to. If I'm right this will be a lot of revenue and very little profit for Tesla. At most I hope Tesla would form a new division and then spin it out as a new business at these profit levels.
2) I see little to no way for Tesla to continue to take 30% off the top and leave owner / operators with 2-6% profit margins. If nothing else, some other company that gets a robotaxi into service could decide to take 25% off the top. Then Tesla lowers to 20%. And the race to the bottom is on until, my prediction, Tesla (or other company) is getting 2-6% profit margin and the owner / operator is earning 2-6%. I.e. - roughly an even split of the profits. Clearly with some businesses doing better and some worse.


This might take 20 years or 2 years in individual geographies as its deployed, but I see the outcome as inevitable. Its in the nature of transportation - technological advances and cost improvements flow through to the customers and very little sticks to people buying and deploying the technology because the business is just too big and the competition is too intense for anything else to happen.

I see that in a longer time horizon transportation becomes a Netflix style subscription model with multiple tiers of usage that will even eliminate the fleet owners. Tesla will build centralized robots that will clean dirty cars everyday which is not technically challenging at all. Example - pay $300 a month get 1000 miles on tesla network and do not worry about other aspects such as insurance, maintenance, parking etc. Tesla will still maintain good margins only because the vertical integration will help them lower the costs and be at a scale that one can match easily. There are other high margins opportunities such as having app store for car etc. which provides margin flexibility for Tesla. Basically, it may very well end up looking like Mobile industry that we have now.

 

[Singuy]

1. The Chinese people have no say in this. Xi will decide, given the policies he has undertaken so far I fear he may go for it.
2. China was militarily weak, they were humiliated in 1996 when they sabre rattled and the US sent a carrier between China and Tawain and they realised they couldn’t do anything. So they started a modernisation drive (which now ironically they could fund as a result of opening up the economy in the 1980s) Now that modernisation is paying off.

I don’t want to get to much into the possible reasoning behind a potential war other than to inform the shareholders on this board to factor this into account and to watch this closely because if a war does start it will probably be a surprise attack and straight away the stock will tank. Irrespective of what the US does in response.

Look, every country can't hold a candle to the U.S military so it's stupid to even try. Every country also can't deny the fact that China is the heart of world manufacturing.

China and the US starting a war against each other is a loser for humanity. The crash in the market will make Covid's seem like child play but not that money even matters since hyperinflation will take place after trade stops.

I believe every country understands what they want and just needs to maneuver through it. The US or WH must look like they are not soft on China so they must poke the bear. The Chinese must look as if they can't just be pushed around as a showcase of strength to their people. They will play this game until Chinese's face is restored and the U.S starts dealing with some other post Covid crisis that doesn't involve the Chinese and the world continues to prosper.

 

[JustMe]

Good god will this weekend never end?

 

[adiggs]

I see that in a longer time horizon transportation becomes a Netflix style subscription model with multiple tiers of usage that will even eliminate the fleet owners. Tesla will build centralized robots that will clean dirty cars everyday which is not technically challenging at all. Example - pay $300 a month get 1000 miles on tesla network and do not worry about other aspects such as insurance, maintenance, parking etc. Tesla will still maintain good margins only because the vertical integration will help them lower the costs and be at a scale that one can match easily. There are other high margins opportunities such as having app store for car etc. which provides margin flexibility for Tesla. Basically, it may very well end up looking like Mobile industry that we have now.

Well that is certainly an interesting idea - transportation on a monthly subscription. I already see that coming to energy (electricity).

That'll need some more thinking

 

[asburgers]

Well that is certainly an interesting idea - transportation on a monthly subscription. I already see that coming to energy (electricity).

That'll need some more thinking

That's what CANOO is attempting.

 

[buttershrimp]

I see Tesla launching its own app... Tuber. I have to tuber to the airport. Want to share a tuber? My tuber smells like cabbage. Try it.... it’s great fun!

 

[StealthP3D]

10x better than a human will eliminate all "stupid stuff". Basically the car will end up in an accident after drunk drivers and texters runs into a Tesla.

So we will see what happens in the court of public opinion after waymos first accident.

Most accidents shouldn't happen. That doesn't mean it's not good to prevent them.

I will be monitoring FSD (with my butt in the driver's seat) until it's at least ten times safer than your average driver and long after it has already reduced the overall accident and death rate. They say "two heads are better than one" and a FSD car being monitored by a human is safer than either one alone.

 

[Singuy]

Most accidents shouldn't happen. That doesn't mean it's not good to prevent them.

I will be monitoring FSD (with my butt in the driver's seat) until it's at least ten times safer than your average driver and long after it has already reduced the overall accident and death rate. They say "two heads are better than one" and a FSD car being monitored by a human is safer than either one alone.

10x safer than a human driver may not be provable until all or most cars on the street are autonomous. Elimating your car hitting others will not elimate others hitting you. A good avoidance system will help as much as possible but some are physically unavoidable.

 

[[email protected]]

Northbound on I5 today, we counted 31 southbound car transports filled with MYs and M3s as we traveled from Tejon to Patterson. 9 cars on all but a few that had 8. Busy end of q day for Fremont....

 

[StealthP3D]

True, good drivers help make Tesla better over time. But good driver are not plentiful. Why do we assume that random customers are good drivers? A few are, definitely not all. But granted, that is a valid argument.

Tesla doesn't use random customers for training or validating the system. I believe it was Karpathy who let the cat out of the bag a couple of years ago when he quipped they had developed proprietary techniques to filter drivers for special training purposes.

But once FSD is good enough, it will will quickly become better. And then extremely good. We would start seeing a new thing: Handheld OMG video testimonials of Tesla FSD doing crazy sugar in order to get out of really dangerous situations - and doing it succesfully, because it just became that good that quickly.
So I don't see this window of Tesla needing a lot of owners for support being very long. A few months or a year perhaps.

That's a good point. It's likely to become very good at avoiding difficult accidents very quickly. However, I think it's important to realize that the two periods will have years of overlap. FSD will be demonstrating amazing abilities to detect and avoid accidents that humans would stand little chance with while simultaneously doing dumb and fatal things that a human would be very unlikely to ever replicate. Over time, the latter will diminish while the former will gradually become almost perfect.

What matters in the end is whether more or fewer people are being killed and maimed.

 

[Krugerrand]

1. The Chinese people have no say in this. Xi will decide, given the policies he has undertaken so far I fear he may go for it.
2. China was militarily weak, they were humiliated in 1996 when they sabre rattled and the US sent a carrier between China and Tawain and they realised they couldn’t do anything. So they started a modernisation drive (which now ironically they could fund as a result of opening up the economy in the 1980s) Now that modernisation is paying off.

I don’t want to get to much into the possible reasoning behind a potential war other than to inform the shareholders on this board to factor this into account and to watch this closely because if a war does start it will probably be a surprise attack and straight away the stock will tank. Irrespective of what the US does in response.

Ok, so what do you think investors should do? At least the Covid-is-going-to-end-the-world people yelled sell at the top of their lungs, while selling their holdings.

Telling us something might happen in the world to cause the SP to decline, but providing zero semi-solid guidelines - like dates - is unhelpful since this is what we live every single day.

Yes, I find dramatic warnings tiresome because they’re often irrational and overdone and emotional.

Ok, fine. I’m factoring in your warning of a potential war and my conclusion is; go back to playing dead and continue to hold.

 

[TheTalkingMule]

Sounds like the Friday mystery selloff gas been solved.

Goldman U-Turn on Hwang Put Bank at Nexus of Margin Call

Bill Hwang, a former hedge fund manager who’d pleaded guilty to insider trading, was deemed such a risk by Goldman Sachs Group Inc. that as recently as late 2018 the firm refused to do business with him.

www.bloomberg.com

Makeses no sense to me, but I'm glad to hear it's related to one specific shop being margin called rather than some macro headwind.

Let's see $680 at the open to kick the week off right.

 

[buttershrimp]

Does anyone think the gigantic Goldman margin call on Friday caused Tesla shares to plummet specifically? Would we see volume spike?

 

[StealthP3D]

10x safer than a human driver may not be provable until all or most cars on the street are autonomous. Elimating your car hitting others will not elimate others hitting you. A good avoidance system will help as much as possible but some are physically unavoidable.

There is a method to determine who (which vehicle in this case) is at fault. While I see your point it is not something that would prevent statistics from being used to determine when FSD is safe enough to allow on the street. But small uncertainties like this are why the standard will be greater than 1.1 times as safe as a human. Fortunately, the regulators are not strangers to statistical data. They use it very effectively to inform most of what they do on a daily basis.

 

[mblakele]

At the risk of beating a dead horse, I'd like to compare developing FSD to developing a virtual assistant (VA): Siri/Google/Alexa etc. Both require large amounts of user data and AI work, plus a distribution channel.

• VA voice data :: Driving data
• VA speech recognition :: Vision and motion inference
• VA "skills" :: FSD policies and task-specific models
• VA on phones/tablets/smart-speakers :: FSD on cars

Of course the difficulties aren't equivalent. Arbitrarily, developing FSD 1.0 might be 10x harder than developing the first iPhone release of Siri (and developing a human-competitive general AI might be another 10x, 100x, or ???). But the analogy might be good enough to predict the competitive landscape.

VA successes, in no particular order:

• Apple Siri
• Google Assistant
• Amazon Alexa

Failures (so far anyway):

• Samsung Bixby
• Microsoft Cortana
• Facebook M
• etc....

The victors so far have two things in common: AI expertise, and a device platform used to collect data and distribute update. The failures — so far — lack one or both of those.

If we agree that FSD and VA are similar AI problems, and FSD is significantly more difficult than VA... any competition to Tesla is likely to come from a combination of AI expertise and a device platform that can collect data and implement a feedback loop. Contenders include existing AI companies: Apple, Google/Waymo, Amazon; also Intel/MobilEye — but only if they partner with existing transportation platforms or manage to create their own.

Legacy automakers without a strong AI partner risk looking a lot like Samsung and Bixby: trying to fast-follow but unable to get traction.

Looking at FSD through this lens, Tesla combines excellent AI expertise with a strong device platform.

 

[Dutchie]

Am I the only one who is exited for what is going to happen the coming weeks/months due to the hints we were getting about FSD?

 

[Krugerrand]

Sounds like the Friday mystery selloff gas been solved.

Goldman U-Turn on Hwang Put Bank at Nexus of Margin Call

Bill Hwang, a former hedge fund manager who’d pleaded guilty to insider trading, was deemed such a risk by Goldman Sachs Group Inc. that as recently as late 2018 the firm refused to do business with him.

www.bloomberg.com

Makeses no sense to me, but I'm glad to hear it's related to one specific shop being margin called rather than some macro headwind.

Let's see $680 at the open to kick the week off right.

And the shenanigans just keep on keeping on. May they all spend their final days on this planet alone and penniless on a dirty street.