Biggest Tesla FSD Catalyst in 2024 and 2025 Will Be Other Carmakers Licensing

by

If Tesla achieves fully operational FSD with robotaxi capability (before robotaxi regulatory approval) will be like dragons in Game of Thrones. Other carmakers will have no choice but to license it. This will be one of several huge catalysts for Tesla valuation (not financial advice.)

Here are the developments that should positively impact Tesla valuation.

NOTE: On Tuesday, April 2, 2024, Tesla will likely reveal weak first quarter production numbers. There was the eco terrorist fire at the Berlin gigafactory, chinese new year and generally soft demand.

1. FSD off beta [April earnings 75-90% revenue recognition]
2. FSD free trial positive April & signups May
3. FSD 12.4 April
4. FSD 12.5 May
5. Energy Add 1+ GWh Megapack each quarter
6. Change quarter-end volume note adds energy and FSD (April 2? July 2? Oct 2? Jan 2)
7. Change earnings to call out energy and FSD as material. Provide consistent guidance. Force analyst recognition
8. First OEM license FSD/software [April-Aug?]
9. FSD Feature Complete
10. FSD 12.5+ quality/comfort drives major adoption increase
11. FSD drives more car sales
12. FSD rapid learning for quality FSD in China, Asia, Europe
13. FSD starts many local regulatory processes
14. Unsupervised FSD in Boring Tunnel
15. 4680 ramped past 15 gwh per quarter
16. 4680 ramped past 25 gwh per quarter.
17. 4680 ramped past 50 gwh per quarter
18. China megapack factory completed and delivering. Expect sooner than expect because construction is surprisingly fast in China

The biggest possible near-term catalyst would be a series of OEM licenses for FSD. Like the charging station agreements.

Successful FSD with significantly more signups will also break the only a car company narrative. At some point, recategorizing Tesla as a software/AI company will be a huge PE multiple shift.

Auto makers have PE of 10-20 but software applications can have 50+ PE, and with good growth even more.

As of March 28, 2024, (AMZN) price-to-earnings (P/E) ratio was 61. This is based on a share price of $180.14 and diluted earnings per share (EPS) of $2.90 for the trailing 12 months (TTM) ending in December 2023. AMZN PEG (Price Earnings Growth) ratio was 3.44. This was based on its 5-year EBITDA growth rate of 18.60%.

Tesla changing the name from FSD beta 12.X to FSD 12.X supervised makes the eventual easy name change to FSD X.X unsupervised for true robotaxi. Based on regulatory approvals in different places.

15 thoughts on “Biggest Tesla FSD Catalyst in 2024 and 2025 Will Be Other Carmakers Licensing”

    • LIDAR operations like Waymo and Hyundia are limited to tiny hypermapped zones. Waymo with about 500 vehicles in part of SF, Phoenix-Arizona and soon part of LA. Tesla everywhere. The LIDAR and other sensors add about $100k per robotaxi.

  2. All valid questions. The time frame is uncertain. What sets Tesla apart is all the model 3 and model y vehicles feeding information back to the large AI. And Tesla has been working on its weights and balances and bias for years and years.

    Other manufacturers have been working on autonomous driving, but other manufacturers had been working on electric vehicles for decades before Tesla. Tesla does it at mass scale and is profitable. It seems reasonable to believe that Tesla is far ahead of all other manufacturers experimenting with autonomous vehicles.

  3. They won’t be forced to, because it’s not as popular a feature as you assume. Yes, it gets tech nerds excited, because it’s exciting tech. But most people aren’t tech nerds, and won’t put their lives into the hands of machines they perceive to be “worse at driving” than they are. Everyone assumes they’re better than average, so FSD has a long, long way to come before the public’s perception of it is better than their own self image.

  4. Will someone please explain to me how Tesla will be able to stuff petabytes of FSD data into a every car it makes? If they’re going to rely on instant communication back to HQ instead, a single second communications interruption could prove fatal.

    • You are confusing training data with the size of the completed model. Chat GPT was trained on 45 Petabytes. It runs on 500 gb. Granted that 500 GB is ran on a super computer. Optimizations will be necessary. Tesla already does this now.

  5. What evidence has led you to these fantasies of customer and OEM behavior? For starters, why do you believe that Tesla is the only company that has been working on self driving tech? There is GM Supercruise, Mercededs-Benz driver assistant, Nissan ProPilot, Honda, Ford Bluecruise, and VW group and all have level 2 or above systems right now. Ford and VW were using Argo AI, I don’t know if they moved everything in-house or not. Aurora is working with Peterbuilt, Volvo and Toyota on self driving trucks. Do you believe no one at any of those companies has ever heard of neural networks? I can guaranty you that they are all renting GPU time and running all their previously recorded data through newly constructed neural networks themselves. Neural networks mean that the insanely expensive FSD board that Tesla was producing should no longer be needed. That was needed when everything the car encountered had to have individual labels attached to them and those labels linked to responses via percentages. None of that’s needed anymore. There is nothing proprietary in using a neural network to teach a vehicle to drive.
    Secondly, where is the consumer demand? FSD is only purchased by 1 out of 5 Tesla buyers. Supercruise is believed to be under 10 percent. Over 60% of people survived say they are apprehensive about self driving tech. Why would anyone ditch their own research to “bend the knee to Tesla” for a product with almost no consumer demand? I believe the world of self driving tech will be wonderful. Tesla is not the only company with a lot of driver data. I have no reason to believe Tesla has anything proprietary in this area. They posses no “secret sauce” that can’t be replicated by their competitors. They only got here first, and without the consumer demand, its not much of a prize.

  6. What evidence has led you to these fantasies of customer and OEM behavior? For starters, why do you believe that Tesla is the only company that has been working on self driving tech? There is GM Supercruise, Mercededs-Benz driver assistant, Nissan ProPilot, Honda, Ford Bluecruise, and VW group and all have level 2 or above systems right now. Ford and VW were using Argo AI, I don’t know if they moved everything in-house or not. Aurora is working with Peterbuilt, Volvo and Toyota on self driving trucks. Do you believe no one at any of those companies has ever heard of neural networks? I can guaranty you that they are all renting GPU time and running all their previously recorded data through newly constructed neural networks themselves. Neural networks mean that the insanely expensive FSD board that Tesla was producing should no longer needed. That was needed when everything the car encountered had to have individual labels attached to them and those labels linked to responses via percentages. None of that’s needed anymore. There is nothing proprietary in using a neural network to teach a vehicle to drive.
    Secondly, where is the consumer demand? FSD is only purchased by 1 out of 5 Tesla buyers. Supercruise is believed to be under 10 percent. Over 60% of people survived say they are apprehensive about self driving tech. Why would anyone ditch their own research to “bend the knee to Tesla” for a product with almost no consumer demand? I believe the world of self driving tech will be wonderful. Tesla is not the only company with a lot of driver data. I have no reason to believe Tesla has anything proprietary in this area. They posses no “secret sauce” that can’t be replicated by their competitors. They only got here first, and without the consumer demand, its not much of a prize.

  7. Hm, Brian, I will have to disagree. The OEMs will not flock to the FSD for the simple reason that they do not have the FSD HW on their vehicles. And this is legacy OEMs we are talking about… How long for them to install the FSD computer, even if Tesla will supply it to them? That in itself is a couple years for them to install it in a model that i already in production. So no effect on cash flow until, say, 2026? More likely 2027. And this assumes that Tesla will have their FSD breakthrough this year…

    • Are the big earnings will take several years from break through. But stock price will respond well to FSD earnings realization and OEM endorsements when they start. Regulatory world will respond better if Ford & or others endorse/license as well.

    • Further evidence supporting this is that the earliest FSD adopters with Model S/X still haven’t been given recent updates due to hardware issues. Cars never intended for FSD won’t be suitable. Indeed, it will take years.

      • Nope, a Tesla executive said today on X they were working on making software work for earliest S/X, everyone with FSD3 and 4 will get it.

  8. “The biggest possible near-term catalyst would be a series of OEM licenses for FSD. Like the charging station agreements.”

    Great point. This would probably look something like the avalanche of OEMs adopting Tesla’s NACS and finally regulators being forced into recognizing it as the standard. The first would show the pattern of the licensing agreement. That would cause analysts to dust off those old Robotaxi spreadsheets and fill in the numbers to remember why TSLA first got a market cap as high as most of its competitors combined.

  9. Why will anyone pay for it, while it requires supervision?

    Maybe if having it and using it most of the time reduces my insurance rates? Maybe worth $50/mo, more for young drivers?

    While it requires supervision, it seems like a neat gimmick – not a valuable feature.

    Once it doesn’t require supervision, open the flood gates.

    • What about the conscience of the car doing 99% of the driving which has been my experience with FSD 12.3

Comments are closed.