Tesla shares in 2025 is all about FSD, licensing, teslabot and robotaxi

by

I, Brian Wang, made a bunch of correct calls on the stock price movement of Tesla over the past few months. Tesla pulled back after hitting a high of $488. This was after more than doubling from $213 at the end of October, 2024. Tesla disappointed with 495000 cars delivered in Q4 2024, when there was expectations of 506,000 cars. There was expectations of about 8-9 GWh of megapack energy deliveries but Tesla beat that with 11.2 GWh. The energy outperformance will be mean more profit than the lower profit for the cars.

I believe that auto margins will will be over 20% when earnings are reported.

However, the most important part of the Tesla stock performance is fast improvement with FSD (full self driving) both with the AI driving capabilities and with regulatory approval in China and Europe, getting to unsupervised robotaxi and FSD licensing by other car companies.

9 thoughts on “Tesla shares in 2025 is all about FSD, licensing, teslabot and robotaxi”

  1. There is a general issue about FSD (and all the automatic driving tech so far): they do not learn from you and from the roads you regularly drive on. Your driving data are used for the general training of the newer version together with million other users so your specific insights and corrections regarding how to drive on a certain road are completely diluted and averaged. This is the approach of trying to create a good driver that averages all the inputs, but is seeing every road for the first time all the time. I do not know how automatic driving will be solved, but humans work in the opposite way: they might be a below average driver in unexpected conditions, but they might know every turn of a very poorly maintained mountain road and drive quite well there because it is where they live and they learned quickly not to mess up. Obviously the companies want a centralised server farm developing a model and downloading it in every car, but it might be required an approach closer to generative AI where you can distil the information of your drawing/writing style and make a checkpoint specific for you. Stable diffusion has an enormous development community (and leapfrogged DALL-E) because everyone could access and edit the model. Is some major automaker figures out how to make the checkpoints more compact and accessible for an automatic driving system, i and releases them so that the car can learn from the street you drive it will be a gamechanger. only when the car is confident that had learned the road (that specific road) the car will inform the driver that will try to drive itself (under supervision) and depending on the results it might be considered truly autonomous or not. In these scenario Tesla centralised approach to build the super driver in all conditions might become obsolete quite fast.

  2. FSD is still orders of magnitude short of necessary reliability for robotaxis. Latest 13.2.1 looks good, but with critical disengagements still in the 500-1000 mile range (need more data collection for clear picture), they still need to eliminate >99 of of every 100 disengagements to be as safe as a human driver that only crashes every few hundred thousand miles.

    Nothing about current development path suggests that scale of improvement will be possible in next year or two. Much of it will come down to predicting behaviors of other car drivers, pedestrians, children and animals that will almost certainly require far more powerful AI processors almost certainly requires huge hardware improvements.

    I have far more hope for near-term viability of Optimus that robotaxis.

  3. So the share price is all about unreleased and already frequently delayed products. That is worrying. Is Elon throwing his weight around in the UK political scene something to do with trying to promote a political party that will give him an easy ride on licensing his FSD?

  4. Tesla is a meme stock and play thing for swing traders and fanboys. The reality is that pure EVs are not the desired vehicle propulsion between late 2020s and well past the 2050s for the markets that matter. PHEVs and possibly some H2 cell or combustion will likely augment other range-extending systems overwhelmingly in the categories of larger and more utilitarian vehicles that plug-in to wherever is (in)convenient – increasingly not an ‘owned home’. Fossil fuels will likely be in steady production and refinement for decades. Margins on big vehicles matter. China is not the US and never will be in market, culture, or socio-political ambitions, so pretending that what happens there matters in the mid- to long-term is non-sensical. China’s economy is stagnating and its ability to influence world markets is in decline. Witholding critical materials will be China jump-the-shark moment as globalization falters, immigration is suppressed, and supply-line self-reliability takes hold over NIMBYism and reduced need for cheap labor. That being said, automation, AI or not, is still quite a wild card and will likely influence significant exploitation (mining, etc) and industrial sectors — and the big Markets are overbought with interest rates likely stuck for awhile.

  5. Here’s why Tesla fell so hard, according to Statistica (not stock advice): https://www.statista.com/chart/33709/tesla-byd-electric-vehicle-production/
    BYD climbed in EV production and became the #1 EV producer for 2024, while Tesla fell below 2023 and trailed BYD by 4500 vehicles. Yes, Tesla has higher margins, but how long can excessive tariffs protect countries like the U.S. and much of the EU against much cheaper Chinese manufacturers making arguably as good or even better cars in the same price range or lower?

    • The trouble with that theory is that Tesla’s sales actually went up 8.8% in China: https://www.business-standard.com/world-news/tesla-s-china-sales-hit-record-high-in-2024-bucking-global-decline-125010300537_1.html

      It was only outside China that sales went down. If competition with Chinese manufacturers were the main problem then you’d think Tesla would have its worst results in China.

      Also worth noting that a lot of BYD’s cars are plug-in hybrids, not BEVs.

      • The business standard article also says “But global deliveries nonetheless slid 1.1 per cent, missing CEO Elon Musk’s earlier prediction of slight growth. Reduced European subsidies, a US shift toward lower-priced hybrid vehicles and tougher global competition, especially from China’s BYD, were a drag on sales.
        Tesla’s record China sales while its worldwide deliveries fell is reflective of the global EV landscape as China is the only major market seeing robust growth versus a slowdown or even slide in other markets, said John Zeng, head of market forecast for China at London-based consultancy GlobalData.”
        It sounds like a general softening global EV market combined with tariffs on Chinese EVs in America and the EU (Japan too?) are for now favoring Tesla – also since other American, Japanese and European manufacturers are unable to compete with EVs of their own.
        I expect China will depreciate their currency to get past the tariffs with even cheaper exports (and fewer imports) and maybe open a BYD plant in America too. Plug-ins will retain their popularity here (I’ve driven them) due to chronic under-building of charging network; the home-based charging option is getting saturated with people who already own Teslas or want one. Many people don’t have personal garages. Logistical and regulatory challenges will hamper rollout of charging stations, particularly under a Trump administration, notwithstanding Musk’s Best Buddy relationship with Trump which will ebb once Trump attains office and his full power.

Comments are closed.