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

[Cobbler]

What should be considered as a successful autonomy day? What is the minimum expectations you have to call it a succes?

1) overview/presentation/movies of all the FSD functions and timelines?
Personally, I don't think so. What the car should do with FSD is almost 100% known, so it this option would just be a rehash of known things.

2) Real life demonstration where investors can get in the car as passenger (Tesla employee behind the wheel) , driver enables AP from the parking lot and let the car drive itself around the Fremont area with traffic conditions?

I really want investors to quantify the meaning in big $$$ (with ARK pricepoints) of FSD when it's proven Tesla has the cheapest/ most successful way to achieve L4 it with its hw3.0 setup

 

[Cobbler]

I've been staying out of this conversation, but I just want to note is that what Elon said the chips do is health checks, not voting. E.g. you periodically run tests to look for flipped bits or processor errors, and if one core fails its health test, you disable it. The cores don't run the same input all the time and compare the output to make sure it matches; they run different input, hence HW3 gets its capability halved (20x -> 10x) if a core is disabled.
...

No, that's not redundancy. They should process the same inputs and must achieve the same output, otherwise the hardware has been compromised and it's not reliable anymore

 

[Fact Checking]

Second, how many ZEVs would it take to get an FCA/ZEV Pool into complete compliance (91 g/cc)? Plug the numbers into Equation 2 below:

(120)(727,201)/(91)-727,201 = 231,745 ZEVs. Per year.

Ouch again. Obviously, Tesla is not going to get FCA into compliance and eliminate the whole penalty. It can only reduce the pain.

Why not? Tesla has demonstrated manufacturing capacity to make 260,000 Model 3's per year and low capex plans to get to ~400k per year and Fremont capacity to do 500k per year. As per your figures above FCA is facing costs of €2.0b, or about €8,650 per ZEV not made.

If they pay Tesla say €4,000 per pooled ZEV to reduce the penalties to near zero then that's an about +10% margin improvement for European Model 3 sales, and it doesn't burden the Tesla ASP but comes through a side channel. If I was Tesla I'd prioritize European sales big time to both generate income and to demonstrate the ability to bring in the ZEVs.

If Tesla isn't greedy and settles for about 50% of the fine then no matter what FCA management is thinking about paying the competition they won't be throwing a billion dollars a year out the window.

BTW., FCA might also want to partner with Tesla to build more EVs under the FCA brand but with a Tesla powertrain - that would reduce their average emission levels as well. Maybe this is the Tesla powertrain partnership rumor @KarenRei is aware of?

 

[elliott82]

What should be considered as a successful autonomy day? What is the minimum expectations you have to call it a succes?

1) overview/presentation/movies of all the FSD functions and timelines?
Personally, I don't think so. What the car should do with FSD is almost 100% known, so it this option would just be a rehash of known things.

2) Real life demonstration where investors can get in the car as passenger (Tesla employee behind the wheel) , driver enables AP from the parking lot and let the car drive itself around the Fremont area with traffic conditions?

I really want investors to quantify the meaning in big $$$ (with ARK pricepoints) of FSD when it's proven Tesla has the cheapest/ most successful way to achieve L4 it with its hw3.0 setup

I would want them to send two cars on different routes to the East Coast autonomously. That‘ll shut up the naysayers.

 

[KarenRei]

No, that's not redundancy. They should process the same inputs and must achieve the same output, otherwise the hardware has been compromised and it's not reliable anymore

I literally just explained why it's not at all compromised, and how Elon himself pointed out that the system uses health checks, not voting, but okay...

Capacity halves after failover:

"The Tesla Full Self-Driving Computer now in production is at about 5% compute load for these tasks or 10% with full fail-over redundancy"

Health checks disable failed systems. And there's nothing wrong with this at all. I work in a place where most systems are redundant and lives depend on them, and none of them use voting. It's always health checks**. And they're far more sensitive to errors than a neural net; neural nets are by their very nature noisy.

** - There is a difference in that our systems all run in master-slave mode, while HW3 apparently runs both simultaneously until one fails its health check.

 

[RobStark]

O.T Re. Does Advertising Work?

Went to the Supermarket Today. We needed potatoes.

My gal is looking through a stack of 10 lbs bag of russet potatoes of what appears to be 2-3 brands.

I ask her what she is looking for, she says potatoes from Idaho. I look at the bags and they say "From the State of Washington"

She finally finds a bag with potatoes from Idaho. I ask her what the difference is between russet potatoes from Eastern Washington and those from Idaho.

She says "terroir." I *sugar* you not.

 

[Artful Dodger]

This has to be speculation. Stockpiling resources for that long is such an inefficient use of capital.

If they were going to stockpile battery packs I'd expect the to start only a couple of months before production is expected to ramp to create a small buffer. They would obviously be planning increased production of batteries to coincide with the increased production of vehicles.

Couple things:

• of course, its speculation. That's what we do here
• its also about a million bucks interest expense to stockpile enough SR packs for one quarter's production @ 3K per week.

Revenue from 39K base SRs (zero options, no EAP/FSD) would be $1.365B

Potential Gross margin up to $273M USD.

Then add in your margins for upgrades (trim, wheels, paint). They'll sell the most expensive variants at first until demand ebbs, then roll out less expensive versions.

Then add EAP/FSD revenue at near 100% gross margin (OTA).

That's not an inefficient use of capital at all. Since it enables massive sales in China, its likely to be a 300X ROI for stockpiling battery packs.

Cheers!

 

[KarenRei]

O.T Re. Does Advertising Work?

Went to the Supermarket Today. We needed potatoes.

My gal is looking through a stack of 10 lbs bag of russet potatoes of what appears to be 2-3 brands.

I ask her what she is looking for, she says potatoes from Idaho. I look at the bags and they say "From the State of Washington"

She finally finds a bag with potatoes from Idaho. I ask her what the difference is between russet potatoes from Eastern Washington and those from Idaho.

She says "terroir." I *sugar* you not.

If Tesla finishes expanding into all of the world's major auto markets, and all of them can order all versions of the Model 3, and basic low-hanging-fruit demand levers like towing for Model 3 and a drive unit upgrade for S+X have been done, and leasing has been started, and they're still not hitting their sales targets, then I'll support advertising

 

[Fact Checking]

What should be considered as a successful autonomy day? What is the minimum expectations you have to call it a succes?

1) overview/presentation/movies of all the FSD functions and timelines?
Personally, I don't think so. What the car should do with FSD is almost 100% known, so it this option would just be a rehash of known things.

2) Real life demonstration where investors can get in the car as passenger (Tesla employee behind the wheel) , driver enables AP from the parking lot and let the car drive itself around the Fremont area with traffic conditions?

I really want investors to quantify the meaning in big $$$ (with ARK pricepoints) of FSD when it's proven Tesla has the cheapest/ most successful way to achieve L4 it with its hw3.0 setup

I agree that #1 alone won't be considered success - it would just be warmed up hype. But even if it's just a presentation there might be some new, interesting pieces of technical information volunteered, since Elon, Andrej Karpathy and Pete Bannon will be there as well.

I'm reasonably certain we'll see #2, i.e. we'll see real life demonstrations of the new HW3.0 neural networks with all sorts of not yet demonstrated FSD features like stop sign and traffic light recognition.

I'd consider the Autonomy Day 100% success if they also demonstrate left turns and right turns in city traffic as well.

Bonus points for any demonstration of Tesla Network functionality...

 

[KarenRei]

I agree that #1 alone won't be considered success - it would just be warmed up hype. But even if it's just a presentation there might be some new, interesting pieces of technical information volunteered, since Elon, Andrej Karpathy and Pete Bannon will be there as well.

I'm reasonably certain we'll see #2, i.e. we'll see real life demonstrations of the new HW3.0 neural networks with all sorts of not yet demonstrated FSD features like stop sign and traffic light recognition.

I'd consider the Autonomy Day 100% success if they also demonstrate left turns and right turns in city traffic as well.

Bonus points for any demonstration of Tesla Network functionality...

I've seen some videos of the latest deployed release doing left turns. Far from a 100% success rate, but that's starting to go live even on release versions without HW3, let alone development versions on HW3.

I'm actually starting to get excited about this event. My initial reaction was "meh".

 

[vigleik]

So you wanted the 80-90k range? You would have preferred half a billion in extra FCF losses?

I think you make many positive contributions around here, but this type of reply is...not helpful. It might even contribute to making TMC an echo chamber.

I was also hoping for higher production of Model 3s in Q1. As well as higher deliveries, yes. But I would also have been happy with another 10K produced and loaded up on ships headed for Europe and China. I think we can agree that there’s sufficient demand.

You seem to think that Tesla chose not to produce more cars than they did. While that is undoubtedly true for S/X, I think they made as many 3s as they could and there were just too many challenges this past quarter to ramp much more.

 

[KarenRei]

I think you make many positive contributions around here, but this type of reply is...not helpful. It might even contribute to making TMC an echo chamber.

I was also hoping for higher production of Model 3s in Q1. As well as higher deliveries, yes. But I would also have been happy with another 10K produced and loaded up on ships headed for Europe and China. I think we can agree that there’s sufficient demand.

You seem to think that Tesla chose not to produce more cars than they did. While that is undoubtedly true for S/X, I think they made as many 3s as they could and there were just too many challenges this past quarter to ramp much more.

No. I'm very happy that Tesla cut back the S+X workforce to save money and made fewer total vehicles this quarter, rather than stocking up on inventory / in-transit vehicles and adding half a billion dollars in negative FCF and continuing to spend money at Q4 levels.

I think you'll find that the overwhelming majority of Tesla investors are happy about this as well.

 

[schonelucht]

Would the pooling agreement have to be filed as an 8-K at all?

Legally, I don't know. I still think that payments this quarter under the pool (if any) will be minimal.

 

[Prunesquallor]

Why not? Tesla has demonstrated manufacturing capacity to make 260,000 Model 3's per year and low capex plans to get to ~400k per year and Fremont capacity to do 500k per year. As per your figures above FCA is facing costs of €2.0b, or about €8,650 per ZEV not made.

If they pay Tesla say €4,000 per pooled ZEV to reduce the penalties to near zero then that's an about +10% margin improvement for European Model 3 sales, and it doesn't burden the Tesla ASP but comes through a side channel. If I was Tesla I'd prioritize European sales big time to both generate income and to demonstrate the ability to bring in the ZEVs.

If Tesla isn't greedy and settles for about 50% of the fine then no matter what FCA management is thinking about paying the competition they won't be throwing a billion dollars a year out the window.

BTW., FCA might also want to partner with Tesla to build more EVs under the FCA brand but with a Tesla powertrain - that would reduce their average emission levels as well. Maybe this is the Tesla powertrain partnership rumor @KarenRei is aware of?

I was trying to give a realistic example for 2019 and see if I could understand the "low hundreds of millions of euros" statement. But you're right. I’m actually all for it for the reasons you give and I mentioned in a subsequent post that I thought this had the potential to distort Tesla export strategy. (Note Tesla would have to sell 230,000 ZEVs per year in the EU.)

Given the brutal nature of the EU CO2 penalties, almost any manufacturer would kill if they could effectively add a couple hundred thousand ZEVs to their fleet. BTW, if I read the rules correctly, it looks like post-2025 there are direct relaxations in the fleet CO2 caps if the fleet incorporates a significant number of ZEVs, but I have no idea whether that applies to pools or not. Every 1% relaxation in FCA's cap would save them ~€70M/yr.

 

[Oil4AsphaultOnly]

Couple things:

• of course, its speculation. That's what we do here
• its also about a million bucks interest expense to stockpile enough SR packs for one quarter's production @ 3K per week.

Revenue from 39K base SRs (zero options, no EAP/FSD) would be $1.365B

Potential Gross margin up to $273M USD.

Then add in your margins for upgrades (trim, wheels, paint). They'll sell the most expensive variants at first until demand ebbs, then roll out less expensive versions.

Then add EAP/FSD revenue at near 100% gross margin (OTA).

That's not an inefficient use of capital at all. Since it enables massive sales in China, its likely to be a 300X ROI for stockpiling battery packs.

Cheers!

The potential ROI is nice, but maintaining the health of the batteries without an active BMS is just asking for more trouble than its worth. Besides, stockpiling battery packs is just not as capital efficient as bringing up another battery cell line in time to supply GF3.

 

[AlexS]

What should be considered as a successful autonomy day? What is the minimum expectations you have to call it a succes?

1) overview/presentation/movies of all the FSD functions and timelines?
Personally, I don't think so. What the car should do with FSD is almost 100% known, so it this option would just be a rehash of known things.

2) Real life demonstration where investors can get in the car as passenger (Tesla employee behind the wheel) , driver enables AP from the parking lot and let the car drive itself around the Fremont area with traffic conditions?

I really want investors to quantify the meaning in big $$$ (with ARK pricepoints) of FSD when it's proven Tesla has the cheapest/ most successful way to achieve L4 it with its hw3.0 setup

Unless there will be: one more thing. Expect short attack and decline of 10% like on every other Tesla event in past several years.

Well, it was kinda joke. I expect they will impress.

 

[Artful Dodger]

Porsche 911 Owner's First Drive Impressions Of The Tesla Model 3 ("I WAS WRONG...")

 

[Mars ☰mperor]

I agree that #1 alone won't be considered success - it would just be warmed up hype. But even if it's just a presentation there might be some new, interesting pieces of technical information volunteered, since Elon, Andrej Karpathy and Pete Bannon will be there as well.

I'm reasonably certain we'll see #2, i.e. we'll see real life demonstrations of the new HW3.0 neural networks with all sorts of not yet demonstrated FSD features like stop sign and traffic light recognition.

I'd consider the Autonomy Day 100% success if they also demonstrate left turns and right turns in city traffic as well.

Bonus points for any demonstration of Tesla Network functionality...

Since it's going to be a webcast for everyone, I guess there is more likely going to be something like a presentation instead of a demonstration?

Don't get me wrong, I would love to see real environment demo in real time with investors inside a Tesla.
None of us wants the real time demo to go wrong in front of the whole internet. But a demonstration just can not be as easily choreographed as a presentation.

 

[vigleik]

No. I'm very happy that Tesla cut back the S+X workforce to save money and made fewer total vehicles this quarter, rather than stocking up on inventory / in-transit vehicles and adding half a billion dollars in negative FCF and continuing to spend money at Q4 levels.

I think you'll find that the overwhelming majority of Tesla investors are happy about this as well.

I thought the context was that @Dutchie and @sparcs were unhappy about Model 3 production. I completely agree that given the demand situation for S+X it was prudent to cut back, and I don’t see anyone disagreeing with that. (Of course some argue that Tesla should have done a refresh, but that’s a different discussion.)

Do you think it would have been better or worse if Tesla had produced an extra 10K Model 3s in Q1, with those extra Model 3s now in transit?

 

[Singer3000]

If Tesla finishes expanding into all of the world's major auto markets, and all of them can order all versions of the Model 3, and basic low-hanging-fruit demand levers like towing for Model 3 and a drive unit upgrade for S+X have been done, and leasing has been started, and they're still not hitting their sales targets, then I'll support advertising

A genuine question, what is it about the advertising industry that concerns you so much?

Why are you so sure that direct leasing is a superior demand stimulant under all circumstances? It carries an upfront cashflow penalty for one thing. Why are you so certain that advertising in existing large markets is a worse idea than opening up in new and potentially challenging markets, with the SG&E that comes with that (e.g. unscalable service infrastructure, management overstretch, fresh jurisdictional risk and analysis, other idiosyncratic cultural or regulatory challenges etc...)? Or that the R&D spend and downtime in production to upgrade the motors for a marginal performance improvement is definitely worth it?

Why do I get the feeling that if Tesla launched some low key targeted advertising next month, you’d be the first to say what a genius Elon is for doing it and why it was such an obvious move?