Tesla Q1 2024 Earnings Call

Tesla’s Q1 resuts were weak which was already expected given the earlier announment of vehicle deliveries. They did say they expect Q2 to be a lot better given that they had to deal with so many unique challenges in Q1 (shipping through the red sea, power sabotage at the Germany plant, etc.)

But on the call, most everything I heard gave me increasing confidence in the future of this company and as an investment. The market seems to agree as the stock is up about +11% after hours…granted the stock has sunk to a very low level in recent weeks so it is still well below where the shares traded not that long ago.

Keep in mind I already own a good sized stake in TSLA and I fully expect it will be a multi-year investment as their various areas of new business play out

I would highly recommend anyone that owns the stock or is considering it, listen to their earnings call. It is a lot more listen-able than many other earnings calls. They don’t waste your time (or theirs) re-reading most of the same information that is already in the press release. Elon makes some brief comments upfront and then they start taking questions, several of which are submitted ahead of time by normal investors.

A few notes I jotted down while listening (note that the “quotes” may not be verbatim but should be very close to what was said)

Tone

Ovearall, Elon seemed very happy and upbeat throughout the call, even cracking a few jokes. It was a 180 degree contrast to the way he sounded a few quarters ago when just before the first cybertruck deliveries.

New Lower Cost ($25k) Model

The rumors that Tesla would be abandoning a lower cost model automobile appear to have been unfounded. They actually said they are moving up the expected rollout of a new lower priced model, originally estimated for the second half of 2025, to most likely occur in the first half of 2025. They say they are able to move up the production by utilizing the existing configuration of production lines for the new model, rather than having to reinvent the production line as had previously been contemplated.

Cybertruck prouction has just reached 1,000 units per week in April, which I believe is better than most estimates.

Competitors and Supplier Pricing

Elon stated that Tesla believes that vehicle battery orders from their competitors have been dropping dramatically recently. Tesla therefore is currently getting much better pricing as a result of the suppliers’ excess capacity.

He did go on to say that this disconnect between supply and demand is likely to correct itself before too long, but will have a positive impact on their cost of goods in the near term.

Reorganizing the Company Structure

They didn’t provide many details but touched on the company reorganizing itself as it prepares for future areas of expected growth. One item they did mention is that Elon believes it has become way too complicated and time consuming to simply order a Tesla vehicle. He thinks they can and should get to a place where a person can order a Tesla car in “under a minute”.

Autonomous Driving and Robotaxi

Elon made what I thought was a great analogy comparing future autonomous driving and robotaxi to the way that elevators evolved in the past. There was a time when all elevators had an operator that was responsible for moving the elevator from floor to floor and opening and closing the door at the right time. He noted that sometimes the operator could be tired or not feeling well, or simply not as focused as they should be and it could result in people getting injured if the door opened at the wrong time or the elevator moved when it should not. But today, you just press a button and “it just works”. You don’t have to worry about doors opening a the wrong time or the elevator moving when it shouldn’t. We trust that it is programmed to do the right things at the right time.

Of course there are many more variables in a vehicle operating autonomously, but the Tesla team expects that there will be a day in the not too distant future where people summon a car from their phone and it takes them where they need to go and “it just works”. Elon joked that they are “putting the auto in automobile”.

They expect that they are going to manage and run the robotaxi fleet and said to think of it as a combination of Uber and Airbnb. The uber side relates to vehicles that Tesla will own and use in the robotaxi fleet, utilizing them as much as practical. They will also enable people that own Tesla vehicles to opt into allowing the vehicle(s) that they own to be used for robotaxi rides when the owner does not need to use the car. The owner would be able to click a button on the app whenever they want the car to go into service (and I’m sure the owner would collect fees for the car’s use as a robotaxi) and then they could summon the car back whenever they want to use it again (I’m sure once it has completed any in-progress ride). This is the airbnb aspect where they compared it to someone renting out their home or apartment on Airbnb when they don’t plan to use it and not rent it out whenever they want to use it themselves.

They aren’t concerned about regulations and laws allowing full self driving when the system is ready. They mentioned that other competitors (e.g. Waymo) had already been working through some of this which is very helpful to Tesla. They commented that if the data shows that fully autonomous vehicles get into accidents and result in injuries and fatalities at significantly lower rates than human driven cars, then regulations preventing them from operating in certain jurisdictions will result in accidents and injuries and they fully expect, by the time, they are needed, it will not be a major hurdle to get local cities on board.

Licensing of Full Self Driving

Elon said they are currently already in discussions with “one major automaker” regarding licensing FSD, which I believe is new news and really good to hear as this would be a very high margin, potentially very profitable line of business if they can get a few big automakers to license the self driving from Tesla.

He later said there is a “good chance we sign a deal this year. Maybe more than one” but he went on to describe that, even when FSD is ready, it is estimated it could be 3 years before they really start using it given the time it will take for the cameras to be integrated with the licensee’s vehicle models.

Use of Idle Excess Vehicle Compute

There was an interesting discussion, which later went on a bit further after an analyst follow up query at the very end of the call, where Elon noted that he is thinking about ways to potentially utilize and monetize compute power within worldwide Tesla vehicles when the cars are at rest parked or being charged. This, I imagine, would be similar to the airbnb situation above for loaning your car to Tesla’s robotaxi fleet, but not for ride share services, but instead to opt in your car when not using it, for the monitization of the excess compute power in the car (assuming there is enough electrical power charge that won’t be needed for driving right away).

He said, as the number of Teslas on the road expands (he is envisioning there being “tens of millions” of them out there in not too many years and he’s thinking there will be much more compute power in the vehicles in coming years than already exist in today’s models), it would be a waste of a huge amount of usable compute power to just have them sit idly in a garage for so many hours at a time.

Elon compared it to Amazon’s AWS example, where they originally built their servers for a particular use, mostly for use with their own business, and eventually realized they had, and could create more, excess compute which could be sold to others and is now the most profitable part of their business. So he thinks about whether it’s realistic for the excess compute power in the vehicles to be monetized by doing something useful. He wondered aloud if there may be a day when the combined idle compute power of Tesla vehicles parked could exceed more than the total compute power of any individual company.

Obviously I wouldn’t put any value or high expectations that this will become a reality anytime soon. I suspect people would even be wary of their car’s computer being used for who knows what could cause it to get a virus and become unusable when they need it, etc. But’s it’s interesting to hear the thought process and the types of things being considered as ideas for the future.

Optimus Humanoid Robots

Elon reiterated what he has said in the past that he thinks Optimus could one day be more valuable than the rest of the Tesla company combined. He estimated that they could have Optimus robots performing productive tasks in their own factories by the end of this year and may be ready for external sales by the end of 2025.

Wrap Up

That’s all I noted. I totally get that this company is not going to be for everyone. There are mountains of execution risks in order to get any of their growth areas to generate the type of profits that I’m hoping for. But I do think this company will continue to change the world and find ways to do it profitably. I don’t know how much is going to really happy over the next 12 months or so, but I know I want to be an owner when it does, and I expect the story will be great to follow over at least the next 3 to 5 years, and probably well beyond.

-mekong

61 Likes

Thank you for summarizing mekong.

They say you should not add to your losers, but I’ve been averaging down in TSLA. I’ve also been thinking a lot more seriously about buying one in the next year. I’m still feeling very positive about the future.

I read a tweet that said they currently have 35,000 H100 training GPUs installed - with 85,000 coming by EOY. (I’m not sure if that is 85k total - or 85k additional)

Recent layoffs will save them $1B/year.

Cybertruck has ramped to 1,000 per week.

Tesla will look at buybacks once FCF ramps.

They did not sell any of their Bitcoin last Q. (Reminder that if you are a TSLA investor, you are also partially invested in Bitcoin)

24 Likes

$TSLA after hours is up 38% since January 6, 2023.

Not that it matters, it’s just to show how silly it is to cherry pick dates. Stick to fundamentals, specially at Saul’s!

The importance of this earnings call is the step change at Tesla from hardware (Decreasing Returns) to software, AI (Increasing Returns).

As a life long programmer I can confirm the difficulty, in fact the impossibility, of solving the self driving problem with heuristics. From a programmer’s point of view, there are two human brains, the lizard brain that we share with all other animals, and the boolean or reasoning brain which, while not exclusive to humans, is present in few other species and less developed.

Heuristics is done by the boolean brain. It needs to produce definite outcomes for all cases, when no such outcome is possible you get an error code. We have managed to get around this limitation with algorithms like Linear Regression which is used to predict the value of a variable based on the value of other variables. When eight or nine cameras feed data to a computer there is no “variable” to input for Linear Regression to work on, all there is is a stream of pixels. The lizard brain developed to deal with this kind of data. The shortcut term is ‘Pattern Recognition.’

Again as a programmer, not as a biologist, the lizard brain stores all the data it collects over its lifetime. Under hypnosis people can recall much of their lifetime experiences. One of the functions of the lizard brain is the fight or flight response. It doesn’t need exact variables, only which of the two seems to have the better outcome by comparing the incoming pattern vs. stored patterns. When overtaking a car on a two lane road and you see another car approaching head on, how do you decide to hit the brake or the acceleration. Give this situation a lot of thought, this is a problem the lizard brain deals with.

The above explains why Elon Musk and Tesla are rotating from the affordable mass market EV to the RoboTaxi. FSD v.12 is proof that autonomous driving is not just feasible but practical. The rate of improvement (learning) of neutral network based v.12 is orders of magnitude faster than for the previous heuristic based versions.

From an investor’s point of view FSD is Increasing Returns, each new customer costs very little, while cars are Decreasing Returns, each new car cost thousands of dollars to produce.

Denny Schlesinger

o o o o o o o o o o o o o o o o o o o o

Critics says that Large Language Models generate lots of plausible nonsense as proof that AI does not work. They miss the point entirely! How does the Scientific Method work? Nobel Prize winning theoretical physicist Richard Feynman, explained it beautifully, “You start with a guess…”

The guess is generated by the scientist’s lizard brain! Then the guess is tested with experiments, the work of the boolean brain. The function of the Scientific Method is to separate the wheat from the chaff output by the lizard brain!

33 Likes

Some additional information around the robotaxi that I think is interesting and pertinent to investors

First, I didn’t catch this on the earnings call, but it sounds like Tesla is referring to it as the “Cyber Cab” and that will potentially be the official name (not “robotaxi”) of the app and service.

Some screenshots from the development of the app were shared which gives me the impression that a service is pretty far along and may even be launched before they have fully autonomous driving.

Here is the image that Tesla shares:

If that doesn’t show up on my post, you can click this link to see it and this page makes it easy to zoom in since it’s somewhat hard to see the details without zoom given that the background is dark too

It looks like it allows you to set the temperature before the car arrives so that it can start adjusting before you get it. You can also use your own phone to control the music while in the car.

If someone wants to opt-in their vehicle for the robotaxi/cybercab service for a few hours, they can also choose to only make it available to their friends & family, or only to highly rated customers, etc.

One of the unofficial blog conversation that I was following online (not from Tesla) was discussing whether Tesla plans to launch the service soon (this year) more like Uber/Lyft initially where there will be a paid driver in the driver seat.

If they did this, I imagine they would encourage (or require) the driver to keep the vehicle in self driving throughout the ride (probably except when an urgency requires them to take control).

This sounds, to me, like a great idea:

  1. it woud give the company a lot more real-life situation data of how Cyber Cab rides would go in actual situations, which will not only help with the service, but also continue to train the autonomous driving overall

  2. having a paid (uber/lyft style) driver in the car, while the automobile is operating in Cyber Cab self driving, I imagine the driver would have to give Tesla feedback of what works and doesn’t work after each ride, explaining any problems with the passenger pickup, or dropoff, or anywhere in between when the car is operating autonomously as a cyber cab, and what situations caused the driver to need to take control to ensure a relatively smooth rideshare experience.

To me, this would probably be extremely helpful and maybe the most valuable part of launching the service with drivers supervising the ride service initially

  1. It would also start to get customers used to riding in a cyber cab style self driving vehicle. People would initially take rides while there is a driver there ready to take control if needed. So this would likely make those same customers more comfortable when the day comes that Tesla can switch over to fully (no driver) autonomous cyber cab services, because they have already experienced rides where the driver was there, but didn’t have to touch the wheel or pedals and the car got them where they were going just fine.

All of this just makes me more and more optimistic that this isn’t some far off dream and is getting closer, and we may start to see a non-FSD version in the real world very soon, potentially not long after Tesla’s scheduled August 8th robotaxi event.

-mekong

26 Likes

Looking through the slide deck today and I am NOT seeing anything good. Deliveries are down, revenue is down, margin is down…

I did actually get back into Tesla a little while ago because I expected the Cyber Truck to move the needle and it hasn’t. There is also a slide in the deck showing that the Tesla Semis factory is…BARE DIRT. How is that good news, I wanted more diversity in product.

Yeah, they claim they are making enough new batteries to stay ahead of the truck builds, but they have only delivered ~3000 trucks, so that doesn’t seem like a stretch at all.

So they have dojo, the robot, the energy storage, the charging infrastructure, etc, etc…but seriously none of the “products” for sale appear to gaining traction.

Are we all really here just for the story that some day maybe they will have robo taxies and maybe someday soon the robot will be all the rage, and maybe someday in the future they will make money somewhere…??

I have been marking my earnings dates with color codes for a bit now. Green, yellow, red. Any red earnings and I am out, but those usually happen after a few yellows. I have Tesla marked yellow this quarter, but I kinda want to mark it red and bail out of this company again.

Why are people still holding this? What numbers am I missing?

11 Likes

I have had it red for a few quarters but in my notes I have “the options are unlimited” Which keeps me in it.

Andy

10 Likes

I listened to the conference call just to get an “understanding” of where the business was at. The one thing I took away from the call was how many times Elon said that everyone needs to personally go out and try FSD 12.3. He had to have mentioned that at least 5 times during the conference call and even emphasized it at the end of the call. If you aren’t a believer in what they have done with FSD after that, than don’t be a Tesla investor.

9 Likes

Numbers, especially FCF, are down due to geo-politics messing with Tesla Logistics is my take.

After Q1 numbers, Before Q1 CC (in parentheses are my estimates).

Tesla experienced a 8.5% year-over-year decline in vehicle deliveries in the first quarter of 2024, marking its first annual decline since 2020. The company delivered 386,810 vehicles in Q1 2024 compared to 422,875 vehicles in Q1 2023 (25,000 cars less YoY)

Why?

  1. Everyone got inflation wrong and rate cuts expectations are being increasingly pushed to late 2024 or into 2025 hurting long duration names like TSLA the most. People waiting for lower interest rates, as promised, has led to the highest number, percentage-wise, of used cars on the road in the last 50years!!!

  2. War in Israel causing the Suez canal to become unusable to the global car carrier fleet causing an unprecedent lack of shipping capacity. It’s well documented just google it. Perhaps -10,000 cars)

  3. New Model 3 ramped in Fremont Q1 2024 caused temporary unavailability hurting temporarily sales but was misunderstood by investors, perhaps -30,000 cars not sold just due to this.

  4. Cyber Truck ramping issues, not expected? There has been one reguiring a rivet needing to be added to the gas peddle. This slowed production a bit; but, still Cybertruck ramp possibly adding conservatively near 10K units in Q2 (Q2 has 91 days, 13 weeks, implying a needed CT production rate of ‘just’ 769/week). This appears highly achievable due to Tesla reporoting “over 1,000” Cybertrucks “in a single week” already.

  5. Tesla Giga Berlin had to shutdown due to part shortages from Houthie attacks in the Red Sea making shipping routs much longer(perhaps -20000 cars not sold)

  6. Berlin had to be shut down a second time in the same quarter after an arson attack. -2000 cars.

Excess Q1 in-transit inventory, due to logistics not demand(!), arriving EARLY in Q2 from logistical issues

On the Call:

:fire:Tesla will be buying 50,000 additional Nvidia H100s by year-end 2024, increasing its current stock of H100s from 35,000 to 85,000. (Tesla has vastly excellerated the advancement in achieving Full Self Driving- As mentioned in prior post, ”You need to see it to believe it.”

50,000 H100s have a market value (all-in cost) of approximately $2.0B

(This note for those also invested in Nvidia)

My expected exponential growth in Tesla’s current AI Foundational World Model (“no longer compute constrained” and “scalable”, Elon- Tesla increased their number of H100 equivalent compute 130% this last QoQ) could easily lead us to the following situation, I’m not saying it will; but, Tesla is 38% of my portfolio🤩

As mentioned already in this thread, New models possibly in 9-12mo (those of us who read Walter Isaacson’s book recall Tesla spinning up an additional Model 3 line in Freemont, under a tent, in less time).

Tesla using available capacity to bring production from 1.8M last year to 3 Million in the next 12 months could demolish any attempt to continue what IMO is a disingenuous promotion of hybrids by Legacy “Auto” (I believe Legacy is only going to hybrids to explain away their failure in BEVs).

Smart BEVs under $29K (Priced for 85% of US mkt) = “HYBRID KILLERS”

  • FSD lvl 4, many States already have regulations in place.

  • Grok 1.5 is out and very competitive as an AI LLM, Elon said this will be in all Tesla vehicles. Many say he runs too many companies; but, this cross pollination is built in to it.

  • Tesla remains Best in safety, with over a Billion miles driven on Full Self Driving (supervised), with this data, regulators will be known to be killing people if this is not approved.

  • Enabling buyers to join Tesla ride share network, announced on the Q1 CC and Slide Deck.

I agree, If you haven’t Been driven by Tesla’s Full Self Driving V12, you may not be able to understand my conviction.

Closer look at the income statement shows some positives under the surface:

  • Auto gross margin better than I expected (16.4%)

  • Energy margin 24.6%.

  • Tax rate much higher than I expected - numbers would have been better in previous quarters.
    ,
    Combined with bullish news on next gen…no surprise the market likes it. This is very good going into Q2, IMO

Best

Jason

32 Likes

My take on TSLA is pretty simple:
1)Elon has made it clear that investing in TSLA means believing in FSD and AI - and by extension, that means Optimus (the robot).

2)This is a massive bet, with Tesla cancelling the Model 2 program and instead doing somewhat cheaper models, but not going all-in by actually adopting the “Unboxed” production line process they outlined a while back. In Elon’s mind, that’s too much money invested and will take too much time. I think everyone can agree on the “too much time” aspect.

3)There’s debate over what the cheaper models will be - are they actual new models (smaller than 3/Y?) or just cheaper versions of 3 and Y? Given Elon’s hope of shipping them later this year, my money’s on cheaper versions rather than a new model or two.

4)What do you think of this:
Many people continue to wonder if the world’s largest EV seller will ever fulfill its original promise to revolutionize automotive. But now Musk is plotting another new direction …it’s so far from Tesla’s core that you may well wonder if he has finally slipped off the deep end.

Another big idea from Elon Musk? Go ahead and groan. It’s fine with him.

…What it didn’t translate into was the consistent profit growth many investors had expected by now. Lately profits have fallen, dragged down by spending on new technology projects and on FSD transfer offers that Tesla considers marketing in place of TV ads. Analysts expect full-year net income this year to come in at about half of last year’s total. Most worrisome to investors is Tesla’s three-year-plus binge on new technologies.

“I have yet to see how these investments are producing any profit,” gripes Piper Jaffray & Co. analyst Safa Rashtchy. “They’re probably more of a distraction than anything else.”

Neither analysts nor investors think Tesla’s business is in danger of collapse. It’s just that they’re slowly losing confidence in Musk’s promises. The company’s price-to-earnings ratio is much higher than its peers.

What’s more, at the same time Musk is thinking big thoughts, Tesla’s automotive business faces new threats.

Those are quotes from a 2006 CNBC article on Amazon’s announcement of AWS as “a risky bet,” edited by me to change company name, CEO name, and FSD instead of AWS. The article goes on to compare to legacy providers like Microsoft and Google going into the cloud as well.

Maybe the AWS->FSD analogy won’t hold. And TSLA is surely going to continue to come under short term pressure for at least several months. What we’re seeing today shows how quickly investors are ready to dive in on any good news, so this is going to continue to be a volatile E-Ticket ride.

40 Likes

I retired out of manufacturing engineering in the US auto industry.

I have said for a number of years Tesla had not faced one of the costly and negative quality implications of auto manufacturing - effectively managing of volume/model mix/option content fluctuations.

American Manufacturers currently have about 15 labor hours per car. Tesla has over100 man hours per car in spite of all the conversations about enormous automation.

A typical line speed might 60 cars per hour, so each person has 45 to 50 sec of work on car. A 10% line speed change in either direction means every production job in the plant changes by 10%. If the change is an increase, new hires are required increasing the production work force by 10%. The plant layout work station changes for a 10% change will require a two week shutdown, and the line speed will be relaunched initially with only a few cars per day with full speed achieved after a couple of weeks or so. This is the time required for a line operator to attain an ability to meet consistent quality objectives in his/her 50 seconds of work each of the 60 cars per hour.

If the published numbers for Tesla’s man hours per car are accurate, the cost implications of volume fluctuations for Tesla will be substantially greater than American and foreign majors.

I am uncertain of the financial implications of this issue, but to date I have not heard it discussed.

Graydrake

9 Likes

I don’t think they are accurate. Do you have any thing other than your recollection that proves your point? Both in the number of man hours to build an american car and a tesla?

Andy

3 Likes

Indeed, they are not, at least on an apples to apples basis. Most OEMs, especially GM and Ford, are not nearly as vertically integrated as Tesla. For instance, while almost all OEMs buy their seats from a Tier 1 like Magna Steyr, Tesla has a whole group of people making seats.

Also, in terms of hours in the car itself, the then CEO of VW admitted that they take 30 hours per car on the assembly line while Tesla only needs 10:
https://europe.autonews.com/automakers/vw-plans-new-ev-factory-germany-counter-tesla

It is true that Tesla has cuts costs partially by reducing the available options, and it’s hard to argue that this lack of option flexibility hasn’t hurt demand, if not being the primary cause.

In Shanghai, Tesla was rolling Model Ys off the line about one every 45 seconds.

22 Likes

What you described is not how cost accounting works. In brief, the factory is split into ‘cost centers’ and you do the math for each cost center separately. As commented by @Smorgasbord1, Tesla’s vertical integration makes it very different from the more extensive value chain that older OEMs use.

The Captain

13 Likes