My parents have 2 teslas, 2019 MP3, 2024 model y with hw4. Both cars have fsd. They are both in their 70's and have become heavily reliant on FSD now. They will not even consider another future vehicle without FSD, so their next vehicle will be another Tesla regardless of any other feature other car makers may provide (nicer interior, massaging seats, prettier design). They have a ICE minivan, but rarely drives it due to lack of fsd. They are now wanting to eventually replace it with an X or cybertruck so they can have fsd on all their vehicles.
Wow, my wife loves our model Y but I haven’t been able to persuade her to use fsd yet… l use it and can attest to it’s efficacy and safety, w/ still some hiccups though 😅
This tech is already allowing the elderly to maintain their mobility independence as they age and help them drive safer. It will also help those with progressive vision and neurologic diseases and allow them remain independent longer.
One of the many reasons why FSD will explode once it is ready and permitted on streets and roads everywhere. Big revolution in transportation! It's amazing the stock hasn't exploded. Once cybercabs start showing up in numbers, it will. Then will come the other integrated AI and humanoid robots and then more explosions!
I think there is a major hole in the logic of OEM’s adopting FSD in today’s environment. The OEM’s are locked into their hodgepodge of semi-integrated purchased black boxes and lack an ability to build software to integrate those devices (reference VW’s catastrophic attempt to build their own automotive software suite). There would have to be a concerted effort to set open standards across all of the black box suppliers to allow the OEM’s to even attempt to develop their own vehicle software suite, never mind integrating FSD. That just seems to be a chasm to wide to bridge. Tesla has the vertical integration that allows them to make rapid and seamless upgrades (over the air to boot). Given the experience I have had with my CT and FSD V13.2.2, I would love to see it adopted by all OEM’s for the shear leap in road safety, ignoring the profit potential for Tesla and the shareholders. My wife and I were T-boned in a high speed collision by someone who ran a red light. It occurred at a notorious intersection with a problematic entrance ramp that results in its own rash of collisions between the highway and merging traffic. I was focusing on the merging traffic to my right and never saw the vehicle running the red light and approaching from my left. This would have easily been detected and avoided by FSD. When my wife saw the CT in person, she simply said-“I don’t care what you have to do, but I want this truck ASAP!”. It was the best investment I ever made because my wife now feels comfortable traveling on the road again. You can’t put a price on that! 😊😊😊😊😊
Very good point about VW, and especially true when you see how badly they want to merge with Rivian, a very shaky company, just to get ahold of working software for their EV line! That is saying something about how crucial software and AI software especially will be! And only Tesla will have working FSD for the first few years at the least, combined with their ability to supply the massive demand there will be!
Cern thank you for another excellent analysis of a Tesla opportunity I had not considered. Herbert helps to popularize your concepts for all our benefit.😊
Elon said that another manufacturer would have to make a million cars before Tesla would give them FSD. I assume that that is because it is difficult to tailor FSD to a specific vehicle. Look how long it took for the Cybertruck. The manufacturer would also probably have to build the car to be an FSD car. At least cameras but probably lots of other things. So they couldn’t just present Tesla with a bunch of Toyota Camrys or something else popular. They’d have to come up with a new car. Given this, I don’t think a legacy automaker would take the chance of building a million FSD compatible cars, because if for some reason it doesn’t work out while they’re ramping up to a million, that’s a lot of money down the drain. Also, for the cost and time for an automaker to do this, Tesla could do it for less cost and time. They could add another line, build another factory, etc. This is why I don’t think other automakers will ever get FSD.
Great point! They are truly going to be stuck. Let's see how hard they quietly lobby against federal approval for FSD. If Tesla get a fed green light for public roads and highways, that will be huge!
I love FSD v13.2.2 and think a newer version of FSD will change the world very soon. But, there are tiny nuances that will take a bit to solve. These are my experiences (not throwing shade on anyone or anything, just things to work on) . Please share your experiences too or if you have experienced similar things. 1. Respecting speed limit - slow down even if other traffic is going faster. The whole idea is keep within the confines of the law. A simple audio command like "respect the speed limit" should make it slow down. But it should do it always regardless. 2. Inaccurate maps for superchargers ( around 540 miles of driving) - Randomly turning in the opposite direction of super charger locations (happened 3 times yesterday out of 5 stops and once day before) . I presume the maps that Tesla has are not accurate in terms of geo locations. Or something else. 3. Maps again - Does not stop at the right location based on the address and randomly parks in the middle of parking lines or parks ahain in the middle of crossed line parking areas or handicap location. Could be a general Tesla maps issue or something else. I would think it could figure out a line in the parking area. 4. Turning into a lane ending in 1000 ft on the highway. It did that 2 times yesterday. I had to intervene once since there was car to the right. 2nd time luckily there wasn't and it switched lanes. 5. Some amount of map and traffic intelligence needs to applied while making turns (right or left). It saw a stalled traffic in the left lane about 0.25 miles before a left turn, it turned to right lane and went forward to about 50 yards from the turn and just stopped and signaled left frim the right lane. I had to intervene. FSD needs real time lane traffic to make good decisions and not block traffic for easier flow of general traffic in a particular direction OR make a totally different path by advising the driver that it is looking for a new path due to the traffic. 6. I did notice while on FSD that you can nudge it to change lanes by signaling and slightly turning the steering wheel. Not so much that you exit FSD . Just a tiny bit of pressure. I was happy I could but would love for it to let me do so using a command so it can judge the safety of it as well. 7. Generally driving to certain locations by using certain lane to avoid getting stuck - maybe this can be learned by lane traffic maps. Some way to see ahead and get information so it can make better judgments (this is a generic version of the previous point) I am sure I will find more nuances. But it has only been a week since I got my car. Lots more to explore.
One big challenge is timing of announcements. It will take some time to implement FSD technically, so they are afraid of Osborning their own current lineup. Retrofitting will be hard. On the other hand, the longer they wait to get started, the more they will lose/fall behind.
I love it but I have a question. Does all manufacturers have to pay Volvo a royalty for installing seat belts? I see the government forcing manufacturers to license FSD but will Tesla be considered a monopoly.
Look how desperate VW is to get just working OS software for their EVs that they are spending billions, possibly at a loss, in Rivian. Imagine the desperation over getting working FSD software! Tesla will be getting HUGE offers for licensed software.
You just need to install more smart traffic lights to solve the issue mentioned at 33:40. Those are and have been very common for quite a few years now depending on which city/country you're in. Smart traffic lights simply monitor how much traffic there is and adjust the timing based on that. Only need a few sensors for that; no need to have the lights "talk" to the cars or anything like that.
Great show. There are definitely more revenue streams if OEMs adopt it, some are big enough to offer there own insurance for FSD cars they make. Tesla could clip the ticket on that as well. Everyone is a winner, excluding existing vehicle insurers.
Other OEMs will never license FSD because aside from being slow on design for new models they would need to have some of their supply chain controlled by Tesla (cameras, wiring harness, HW4 control unit, screen). They would need to agree to over the air updates. Unless Tesla becomes like Foxconn and contract manufacturers Fords and GMs. GM already does that in China.
@@oggyoggy1299 Did Nokia license anything from Apple or Samsung to make a better smartphone? No they just went bankrupt. Will the union allow Ford or GM to use Tesla as a contract manufacturer? Never. Imagine how good an electric Ford Bronco would be with 800V architecture, FSD designed and built by Tesla in the US but branded a Ford. Some of the smaller OEMs Rivian, Honda, volvo might license it
It is not as simple as "licensing" FSD. The company, let's say Ford, has to engineer the cameras, wiring, etc etc into their designs. Then mock it up, test it, and plan to build it. That takes a number of years. Best case, Legacy company starts now, and roll out FSD in 3 years.
@@bruce7244 The Fords and GMs won't go out of business in the next 5 years, FSD, or no FSD. For a couple of reasons. 1) Bankruptcy could be a helpful thing. Bankruptcy is not going out of business, it is "financial reorganization". They can get rid of some plants, employees, etc, and it will help to cut their cost. 2) Tesla is not in a position to build the # of cars/trucks/vans/SUVs that Ford and GM build. Maybe in 10 years, but look how long it took to come out with the Cybertruck. US buyers are not embracing EVs yet. 90% of all cars sold in the US yesterday were NOT BEVs. It will take quite a while for buyers to change their perception. Meanwhile, I am very happy with my $TSLA.
OEMs will have to license the whole stack (hardware, communications, software and operating system). To minimize the time, the external footprint dimensions, wheelbase, stering angles, camera locations, etc should correspond IDENTICALLY with a Tesla model.
Not really Waymo’s solution works on ICE vehicles already if it gets to that point they’ll license from Waymo. Whatever Tesla has to offer we don’t know if it will work on an ICE will need 3 years to integrate
@@mandolorian25 Waymo's solution is vastly inferior, Rube Goldberg contraptions costing--what $100 or $200k per copy? That's going nowhere, they may exist for a few more yrs until Tesla completely takes over.
@@mandolorian25 Not really not really... Are you going to pay over $100k just to get Waymo's hardware put onto your car? ICE cars are the Nokias and Blackberrys of the Automotive world. Smart Cars are the thing.
I used to think that if Tesla will ever solve FSD, but now it's a matter of when. Had a chance to try V13 yesterday starting from my driveway for a 35 mins drive. Only 1 intervention was needed (a split road that recently changed from 2 lanes to 1 lane, it did not recognize the new lane change). I was on "Standard" setting and it drives like me, actually it's more smooth than my driving. I haven't had a traffic ticket in over 15 years. Amazing!
Keep in mind for comparison. Here in Europe only a few pay more than 500 Euros in a year for insurance. So the current model has to be completely adjusted if Tesla insurance were to come to Europe
I think the TAM would drop from 90M to 50M vehicles globally as EVs are more durable than ICE cars. Tesla could be dominant in NA and EU with 20M demand.
43:50 exactly this. I would bet big on no licensing for FSD. Y’all have no idea how difficult it would be to adapt FSD to all these OEM platforms. Licensing isn’t just free margin. It will take effort and time to adapt the stack to 3rd parties, and that will slow down FSD development for Tesla cars. When the entire company is riding on FSD getting feature complete, licensing makes 0 sense.
Cern says, "Why wouldn't you want the safest vehicle for your family, so you'd want to buy FSD." I have another way of looking at it. Tesla has always been great at NOT charging for safety related features... Tesla has them coming standard with all vehicles. Take AutoPilot for example, or Smart Automatic Emergency Braking. Other car makers charge for these as options, or at higher trim levels, but not Tesla. So why should Tesla charge for what is obviously a HUGE safety feature? What about the people who can't get Tesla Auto Insurance and who can't afford an additional $100/month or $8000? Is it okay for them have a higher risk of accident/injury/death? What about Tesla having a "safety FSD" feature which comes with the car, and a "chauffeur FSD" feature which you pay for?
I’d like to see the question of how much time would be required for an oem to add fsd to their vehicle. Seems to me it would take a couple of years to do all the wiring and camera additions. Maybe this is a Jeff Lutz question.
If it takes 3 years for an OEM to adopt an FSD license with Tesla vs Tesla using the same 3 years to build a new factory to produce the same number of vehicles, why wouldn't Tesla just expand?
It's absolutely imperative that Tesla either finds a way to release V13 on HW3 cars or offers an upgrade to AI4 so that the 6M cars can run V13. They can charge a small amount, say $250 for that upgrade but without that support, they will not be able to achieve large scale road safety for years until normal or even accelerated attrition eliminates those 6M older cars.
People today pay more than $5k for passive safety systems in a car, like seatbelts, airbags, ABS, shatter proof glasses, traction control, crump zones, roll cages, collapsible steering. None does much to prevent an accident, just protects passengers life’s in case of accident. A single accident avoided by FSD exceeds its cost by large, not reven considering the lives saved. No doubt to me that one day will be mandatory, or at least an option in each car sold.
Same reason it took Biden 4 years to approve the first lithium development in Nevada. And now, they're looking to block it longer because of a fictitious snail.
Question to the cyberbulls/cern: To me there seems to be a pricing analogy between semis/busniss vehicles and cybercabs/robotaxi. After all both are vehicles, that busniss-customers aim to earn cash with. If the above is true: how are semi-sale prices formed, and could the same way be transfered to cybercabs?
We can simplify this. How many CEOs were hired to implement a smart car of the future that will include self driving? Current CEOs were hired to manage the existing business that they were given and this will be "all new business." Wouldn't the first task of an OEM be to think about hiring a new leader to reorganize the company into one capable of making a viable car to sell in this new future? This means working out new suppliers of hardware and of the software. A whole new set of licensing agreements. Realigning all new departments and the whole company. This feels like decades in the making. A brand new company might have a shot at succeeding while a long existing OEM will have to first be willing to remake themselves.
Tesla engineers would be wasting their time trying to teach OEM ICE companies how to integrate the FSD hardware and software suite. Would be like teaching physics to kindergartners. OEMs would basically be building Tesla clones from scratch with Tesla built hardware and have $10K or more added to the manufacturing cost. Easier for Tesla to build more Teslas than to sell hardware to OEMs. Apple briefly licensed Macintosh hardware to others long ago and it was a nightmare.
FSD in CA without safety score vs CO where safety score applies, only saved us $40/mo but could add $50 in a month with a forward collision warning (often false positives).
Omg, my mom is 65 yeas old this year, and she has nerves driven a car in her life simply because she thinks driving could be risky and if something happens, who’s gonna take care of me and my brother? Thus she gave up the thought of owning a car and learning how to drive. Now I am an adult and a long-term TSLA investor, I can’t wait to buy her a TESLA with FSD or even a Robotaxi so she doesn’t have to take bus or walk, her life experience will be changed profoundly, I am so touched by knowing this is going to happen. Thank you guys and Elon😭
we talk a lot about OEM and FSD - what about the car just having only one internal software unlike ford saying 170 vendors parts have separate software.
I don´t know how they do it but for the last 2 years, if you are using satnav to get to your destination, the system can access traffic speed and issues and advise a change of route to avoid the delay on your current route.
None of the legacy automakers have been able to deploy OTA updates and probably never will. For them to develop the hardware for FSD is going to be an impossible task unless Tesla gets involved and sells them the hardware as well. I don't think it is worth it for Tesla to do this hand holding. The big 2 are so used to being bailed out by the government, so they would rather do that. Tesla should go its own way and deploy cybercabs around the world.
I am thinking that the conclusions here were a bit US centric, and missing what could happen in other countries. - FSD may take some time to train for many countries, and each will have regulators that will need to be convinced that it will work on their roads, ditto for Cybercab. - Japan, Korea, Germany, Italy and India may well decide that they need, for strategic reasons to retain control of some auto manufacturing capability, as it feeds directly into defense capability. - EV's not currently useful for people who want to tow large caravans, boats, horses or trailers over long distances, and perhaps the differing form factors of the towed things could make FSD very difficult. - Cybertruck will be far less popular in many places - Japan, much of Europe and the UK for starters - it is simply too big for the roads and parking spaces there.
If Tesla manages to significantly lower their production costs on some models, they could keep the $5K Margin on hardware and include FSD for an extra $5K, preventing buying their cars without FSD. Then a single limitation like the impossibility to drive more than a mile without any passenger inside the car could allow Tesla to ask for a subscription if the owner wants to use their vehicle as a robotaxi (or send it far away to pick someone up). That way the buyers get a cheaper car and have FSD by default to keep them and others around safe, Tesla sells FSD with each car and this can apply to other OEMs using FSD.
IN REGARDS TO THE OTHER GUY QUESTIONING USING OTHER OEMS AT 45:00 IN THE VIDEO. The other problem is the other automakers can not make evs profitably and the question will they ever be able to, is it worth trying to rely on them if there is a chance they will go bankrupt anyway before they can get their robocabs into the market.
I believe Legacy Auto is interested. However as mentioned, it will take a few years to incorporate the cameras and inference chips into their cars to licence FSD, in other words... its very expensive. So they'll likely wait until Tesla has proven the concept by successfully launching unsupervised in a couple of cities. Even if it means paying more by not being the first one to sign on the dotted line, they feel its prudent to wait for confirmation before committing what could potentially be a very large investment. Once proof of concept is in place, they'll flood in. They simply cannot allow Tesla to be the only car manufacturer to have unsupervised FSD because the gap between haves and have nots will simply be too great
Don’t you think CyberCab fleets will be so hyperlocal that the fleet operators are more comparable to fast food franchises than car rental companies? Tesla themselves will be operating the largest fleet.
If there is a need for Mega Packs, I would go ahead to produce more Battery making plants. This can not only benefit Tesla but the consumers to reduce their energy cost plus the environmental solution to cleaner air in conjunction with solar and wind power
Maybe Tesla can license FSD on a quota basis that won't hurt the Tesla owners that wish to be involved in Robotaxi. Why hurt The Tesla owners that have bought a 2nd car just for that purpose. Thanks to Cern and Herbert
To be clear, I don’t think FSD would solve OTA software updates for other OEMs unless they buy all of their electronic components from Tesla. Yes, it would help. But, the real issue is that the firmware on all the various components are built on separate codebases, separate processors, and code languages created by their suppliers. And they have no immediate control over the way the code works in each of these devices. They don’t even have the source code. These devices are just “black boxes”. It’s a big mess.
Tesla FSD would simply not be compatible with those "black boxes". OEMs would basically need to build a Tesla clone with Tesla produced hardware and wiring harness. Not very profitable when Tesla is a competitor.
I have long thought that Tesla will stop making cars for sale, focus completely on Cybercab, drop any idea of licensing FSD and make a fortune for shareholders given the potential profit margins. It's a no brainer and completely meets their stated mission of sustainable energy
Consumers will initially just want to buy vehicles which have the CAPABILITY to run Tesla FSD, with all cameras, etc. The Consumer might not activate the FSD yet, but eventually they will.. So that's it's own potential profit situation than doing an "everyone gets FSD" model. Lastly, what's the cost that these other car manufacturers who are adopting Tesla-inside FSD systems will be paying to license Tesla tech?
I am sure there are others like me that have a Tesla, drive 5-10K miles per year, and do not see a great cost advantage of FSD because 60-80% of my driving is long trips (500 miles each way). Cruise control, which is already included with purchase, does a great job on long freeway drives without the $100/month or FSD price. In addition, if I purchase fsd outright and have a significant accident, there goes my FSD cost. Please give me a constructive argument against my logic. I am retired so daily driving is minimal.
Tesla should 100% NOT share FSD. One big risk is giving it a bad name and bad press through sub-par implementation. Just so other car companies can profit/survive? 👎
Tesla really needs to sell the entire sled to the other car companies. This would allow them to have full self drive fully Incorporated, as well as OTA. Then, all the other car company would need to do is create the body shell to put over the sled.
The ridiculous cost of insurance now day's (thanks for my good driver discount Geico! ugh) will make Robotaxi a more viable alternative to owning a car.
IMO "the Cyber-Cab" revolution will eventually eliminate MOST desire to own cars and all the costs 👺 and liability of car ownership .... And, worrying about theft and dents, and so on and so on...
@@oggyoggy1299 If it costs almost NOTHING... And it takes me door to door... And I dont have to worry about demon-crats breaking into my car while I am inside the store... and I dont have to worry about accident liability... And I can focus on other things on the way there and back... And so on and so on... The answer is YES 😊
I don’t think licensing will happen with any American auto makers. It’s too much for them to install the chips, eight cameras, etc.. The only one might be BYD.
Aren't Ford working on a new EV mass production ready platform in a skunk works headed up by an ex-tesla employee. What if this new ford car has ground up been designed for, and has FSD hardware included already so, FSD ready when it is launched?
I generally agree with CERN, but I would note that no company lasts forever, even Tesla. I know the Ford and Porsche families would disagree, but I think they are wrong. Eventually, a new or better idea, and/or better management will come along. I acknowledge that it may be many years from now, but we should not assume that it will never happen.
Tesla producing 10 million vehicles is going to take a long time. Factories have to be built. It takes time to find sites, get approvals, build the factory, and then ramp it. AND right now there are no new factories on the drawing board. Until demand goes up, Elon is not building anymore factories.
Guys PLEASE Consider this Productisation ….. 2 very different Autonomy s/w packages . 1) ….FSD - as we know it today - at today’s prices or lower …..FSDS (Supervised ) demands seat pressure on drivers side to work : the best ADAS. 2) FSDU - (Unsupervised) no seat pressure demand - allowed to enter Tesla Robotaxi fleet & make money for owner . Different Price but must be related to $ yield :Yield depends upon Miles Driven & therefore must be priced PER MILE Driven using the s/w. Nb Cern’s model misses the Margin that exists on EVERY sale in the channel . (EVERY DEALER WORLDWIDE) Ergo not a viable commercial comparison .
those CEOs whose bonuses are based only on a yearly basis, there is no incentive to think longer term. Licensing and implementing Self Driving will take over a year, it will be 5 years and longer. Plus a lot of effort... that no CEO wants to think about. They will prefer to have "the next guy" to worry about it.
I really doubt that other oem manufacturers could integrate the tesla hardware and software stack into their vehicles. They are really outdated compared to teslas computers on wheels
Adoption FSD can be difficult because especially German car manufacturers have the "not invented here syndrome". They want to invent it themselves and think they can do it better...
I lean toward the scenario where ICE OEMs are targeting selling POA EVs with fsd. But by the time they can bring that to market, CyberCabs have begun to erode POA sales. Tesla is disrupting their own EV sales business. Why bother trying to help OEMs sell into a market that is withering away?
Buying a Tesla? Use my referral link and get up to $2,000 off.
ts.la/herbert23392
My daughter: ts.la/amanda443323
My parents have 2 teslas, 2019 MP3, 2024 model y with hw4. Both cars have fsd. They are both in their 70's and have become heavily reliant on FSD now. They will not even consider another future vehicle without FSD, so their next vehicle will be another Tesla regardless of any other feature other car makers may provide (nicer interior, massaging seats, prettier design). They have a ICE minivan, but rarely drives it due to lack of fsd. They are now wanting to eventually replace it with an X or cybertruck so they can have fsd on all their vehicles.
They are just like me. I am 75 and have a model Y and 3 with FSD
Wow, my wife loves our model Y but I haven’t been able to persuade her to use fsd yet… l use it and can attest to it’s efficacy and safety, w/ still some hiccups though 😅
This tech is already allowing the elderly to maintain their mobility independence as they age and help them drive safer. It will also help those with progressive vision and neurologic diseases and allow them remain independent longer.
My wife and I are in our seventies, have a 2017 Model X and a 2024 Cybertruck, both with FSD. I would not buy a car without FSD.
One of the many reasons why FSD will explode once it is ready and permitted on streets and roads everywhere. Big revolution in transportation! It's amazing the stock hasn't exploded. Once cybercabs start showing up in numbers, it will. Then will come the other integrated AI and humanoid robots and then more explosions!
“Basher one of our favorite Tesla retail analysts” 😊
CERN is the best insight out there!
Thanks
I think there is a major hole in the logic of OEM’s adopting FSD in today’s environment. The OEM’s are locked into their hodgepodge of semi-integrated purchased black boxes and lack an ability to build software to integrate those devices (reference VW’s catastrophic attempt to build their own automotive software suite). There would have to be a concerted effort to set open standards across all of the black box suppliers to allow the OEM’s to even attempt to develop their own vehicle software suite, never mind integrating FSD. That just seems to be a chasm to wide to bridge. Tesla has the vertical integration that allows them to make rapid and seamless upgrades (over the air to boot). Given the experience I have had with my CT and FSD V13.2.2, I would love to see it adopted by all OEM’s for the shear leap in road safety, ignoring the profit potential for Tesla and the shareholders. My wife and I were T-boned in a high speed collision by someone who ran a red light. It occurred at a notorious intersection with a problematic entrance ramp that results in its own rash of collisions between the highway and merging traffic. I was focusing on the merging traffic to my right and never saw the vehicle running the red light and approaching from my left. This would have easily been detected and avoided by FSD. When my wife saw the CT in person, she simply said-“I don’t care what you have to do, but I want this truck ASAP!”. It was the best investment I ever made because my wife now feels comfortable traveling on the road again. You can’t put a price on that! 😊😊😊😊😊
True. They’re reliance on so many suppliers and contracts makes this almost impossible, just as it has made profitable EVs almost impossible for them.
Very good point about VW, and especially true when you see how badly they want to merge with Rivian, a very shaky company, just to get ahold of working software for their EV line! That is saying something about how crucial software and AI software especially will be! And only Tesla will have working FSD for the first few years at the least, combined with their ability to supply the massive demand there will be!
Cern thank you for another excellent analysis of a Tesla opportunity I had not considered. Herbert helps to popularize your concepts for all our benefit.😊
great podcast. great interaction between host and guest. Herbert knows Tesla and Cern is great !!!
Love the thought experiment. Cern is the man!
WOW! Thank You Herbert. Cern's thought experiments & graphs are beyond EXCEPTIONAL. Thank You Cern 🙏
Elon said that another manufacturer would have to make a million cars before Tesla would give them FSD. I assume that that is because it is difficult to tailor FSD to a specific vehicle. Look how long it took for the Cybertruck. The manufacturer would also probably have to build the car to be an FSD car. At least cameras but probably lots of other things. So they couldn’t just present Tesla with a bunch of Toyota Camrys or something else popular. They’d have to come up with a new car. Given this, I don’t think a legacy automaker would take the chance of building a million FSD compatible cars, because if for some reason it doesn’t work out while they’re ramping up to a million, that’s a lot of money down the drain. Also, for the cost and time for an automaker to do this, Tesla could do it for less cost and time. They could add another line, build another factory, etc. This is why I don’t think other automakers will ever get FSD.
There's a chance other OEMs end asking Tesla which are the best car specs in order to tailor FSD in the fastest way. It would be crazy.
Please don't license FSD😢
Great point! They are truly going to be stuck. Let's see how hard they quietly lobby against federal approval for FSD. If Tesla get a fed green light for public roads and highways, that will be huge!
I love FSD v13.2.2 and think a newer version of FSD will change the world very soon.
But, there are tiny nuances that will take a bit to solve. These are my experiences (not throwing shade on anyone or anything, just things to work on) . Please share your experiences too or if you have experienced similar things.
1. Respecting speed limit - slow down even if other traffic is going faster. The whole idea is keep within the confines of the law. A simple audio command like "respect the speed limit" should make it slow down. But it should do it always regardless.
2. Inaccurate maps for superchargers ( around 540 miles of driving) - Randomly turning in the opposite direction of super charger locations (happened 3 times yesterday out of 5 stops and once day before) . I presume the maps that Tesla has are not accurate in terms of geo locations. Or something else.
3. Maps again - Does not stop at the right location based on the address and randomly parks in the middle of parking lines or parks ahain in the middle of crossed line parking areas or handicap location. Could be a general Tesla maps issue or something else. I would think it could figure out a line in the parking area.
4. Turning into a lane ending in 1000 ft on the highway. It did that 2 times yesterday. I had to intervene once since there was car to the right. 2nd time luckily there wasn't and it switched lanes.
5. Some amount of map and traffic intelligence needs to applied while making turns (right or left). It saw a stalled traffic in the left lane about 0.25 miles before a left turn, it turned to right lane and went forward to about 50 yards from the turn and just stopped and signaled left frim the right lane. I had to intervene. FSD needs real time lane traffic to make good decisions and not block traffic for easier flow of general traffic in a particular direction OR make a totally different path by advising the driver that it is looking for a new path due to the traffic.
6. I did notice while on FSD that you can nudge it to change lanes by signaling and slightly turning the steering wheel. Not so much that you exit FSD . Just a tiny bit of pressure. I was happy I could but would love for it to let me do so using a command so it can judge the safety of it as well.
7. Generally driving to certain locations by using certain lane to avoid getting stuck - maybe this can be learned by lane traffic maps. Some way to see ahead and get information so it can make better judgments (this is a generic version of the previous point)
I am sure I will find more nuances. But it has only been a week since I got my car. Lots more to explore.
One big challenge is timing of announcements. It will take some time to implement FSD technically, so they are afraid of Osborning their own current lineup. Retrofitting will be hard.
On the other hand, the longer they wait to get started, the more they will lose/fall behind.
Thank you for your service to the community 🎉🎉
Fascinating analysis. Thanks Basher and Herbert.
I love it but I have a question. Does all manufacturers have to pay Volvo a royalty for installing seat belts? I see the government forcing manufacturers to license FSD but will Tesla be considered a monopoly.
Not if fsd is sold under other branding/companies
I doubt that any legacy OEMs will ever be able to even buy FSD from Tesla.
Another great video discussion! Thanks for sharing your thoughts.
Look how desperate VW is to get just working OS software for their EVs that they are spending billions, possibly at a loss, in Rivian. Imagine the desperation over getting working FSD software! Tesla will be getting HUGE offers for licensed software.
You just need to install more smart traffic lights to solve the issue mentioned at 33:40. Those are and have been very common for quite a few years now depending on which city/country you're in. Smart traffic lights simply monitor how much traffic there is and adjust the timing based on that. Only need a few sensors for that; no need to have the lights "talk" to the cars or anything like that.
Thank You Cern! you're fabulous.
Great show.
There are definitely more revenue streams if OEMs adopt it, some are big enough to offer there own insurance for FSD cars they make. Tesla could clip the ticket on that as well.
Everyone is a winner, excluding existing vehicle insurers.
Great conversation, thanks for your opinions ❤
Other OEMs will never license FSD because aside from being slow on design for new models they would need to have some of their supply chain controlled by Tesla (cameras, wiring harness, HW4 control unit, screen). They would need to agree to over the air updates. Unless Tesla becomes like Foxconn and contract manufacturers Fords and GMs. GM already does that in China.
So they’d rather never have FSD???
@@oggyoggy1299 Did Nokia license anything from Apple or Samsung to make a better smartphone? No they just went bankrupt. Will the union allow Ford or GM to use Tesla as a contract manufacturer? Never. Imagine how good an electric Ford Bronco would be with 800V architecture, FSD designed and built by Tesla in the US but branded a Ford. Some of the smaller OEMs Rivian, Honda, volvo might license it
It is not as simple as "licensing" FSD. The company, let's say Ford, has to engineer the cameras, wiring, etc etc into their designs. Then mock it up, test it, and plan to build it. That takes a number of years. Best case, Legacy company starts now, and roll out FSD in 3 years.
Then they go bankrupt
@@bruce7244 The Fords and GMs won't go out of business in the next 5 years, FSD, or no FSD. For a couple of reasons. 1) Bankruptcy could be a helpful thing. Bankruptcy is not going out of business, it is "financial reorganization". They can get rid of some plants, employees, etc, and it will help to cut their cost. 2) Tesla is not in a position to build the # of cars/trucks/vans/SUVs that Ford and GM build. Maybe in 10 years, but look how long it took to come out with the Cybertruck. US buyers are not embracing EVs yet. 90% of all cars sold in the US yesterday were NOT BEVs. It will take quite a while for buyers to change their perception. Meanwhile, I am very happy with my $TSLA.
OEMs will have to license the whole stack (hardware, communications, software and operating system). To minimize the time, the external footprint dimensions, wheelbase, stering angles, camera locations, etc should correspond IDENTICALLY with a Tesla model.
Thank you 🙏
If Ford Skunk Works isn’t already working with Tesla, they are dead ducks. 🦆
Not really Waymo’s solution works on ICE vehicles already if it gets to that point they’ll license from Waymo. Whatever Tesla has to offer we don’t know if it will work on an ICE will need 3 years to integrate
@@mandolorian25 Waymo's solution is vastly inferior, Rube Goldberg contraptions costing--what $100 or $200k per copy? That's going nowhere, they may exist for a few more yrs until Tesla completely takes over.
@@mandolorian25 Not really not really... Are you going to pay over $100k just to get Waymo's hardware put onto your car? ICE cars are the Nokias and Blackberrys of the Automotive world. Smart Cars are the thing.
Thanks guys… you are doing mind blowing work… you are doing amazing thing
Serious question!! Can you do Fsd on a non EV, gas car ???
Other car makers licensing FSD will want to keep it secret because of the Osborne effect.
You have to sell the fsd hardware too?
I used to think that if Tesla will ever solve FSD, but now it's a matter of when. Had a chance to try V13 yesterday starting from my driveway for a 35 mins drive. Only 1 intervention was needed (a split road that recently changed from 2 lanes to 1 lane, it did not recognize the new lane change). I was on "Standard" setting and it drives like me, actually it's more smooth than my driving. I haven't had a traffic ticket in over 15 years. Amazing!
Just tried it out?! Huh? I have been driving it 100% of the time for months
Omg, I am super looking forward to try it in the near future especially after you guys sharing how amazing it is!
Keep in mind for comparison. Here in Europe only a few pay more than 500 Euros in a year for insurance. So the current model has to be completely adjusted if Tesla insurance were to come to Europe
CERN give us price targets for these scenarios 🙏🏽
I think the TAM would drop from 90M to 50M vehicles globally as EVs are more durable than ICE cars. Tesla could be dominant in NA and EU with 20M demand.
CERN is a genius, all brilliant Stuff Herbert as with Mr Fitz Gerald last night Fantastic listening
43:50 exactly this. I would bet big on no licensing for FSD. Y’all have no idea how difficult it would be to adapt FSD to all these OEM platforms. Licensing isn’t just free margin. It will take effort and time to adapt the stack to 3rd parties, and that will slow down FSD development for Tesla cars. When the entire company is riding on FSD getting feature complete, licensing makes 0 sense.
Byd has this tech (ota) Updates and can tweek the suspension to comfort level and shorten braking distance.
Cern says, "Why wouldn't you want the safest vehicle for your family, so you'd want to buy FSD." I have another way of looking at it. Tesla has always been great at NOT charging for safety related features... Tesla has them coming standard with all vehicles. Take AutoPilot for example, or Smart Automatic Emergency Braking. Other car makers charge for these as options, or at higher trim levels, but not Tesla. So why should Tesla charge for what is obviously a HUGE safety feature? What about the people who can't get Tesla Auto Insurance and who can't afford an additional $100/month or $8000? Is it okay for them have a higher risk of accident/injury/death? What about Tesla having a "safety FSD" feature which comes with the car, and a "chauffeur FSD" feature which you pay for?
I’d like to see the question of how much time would be required for an oem to add fsd to their vehicle. Seems to me it would take a couple of years to do all the wiring and camera additions. Maybe this is a Jeff Lutz question.
Jeff has said it's three years to produce a new car model. Likely longer if they need to redesign to Tesla's advanced hardware suite.
I bought fsd with my new Model y in December. I suspect a lot of people are going to buy it because it's really good now.
If it takes 3 years for an OEM to adopt an FSD license with Tesla vs Tesla using the same 3 years to build a new factory to produce the same number of vehicles, why wouldn't Tesla just expand?
It’s zero auto makers that serious about FSD.
They won’t do it until cybercab is fully working with unsupervised FSD.
V13 is very good I am beginning to question my driving skill.
FSD is my primary driver and I just there to occasionally supervise
Cern is always epic
It's absolutely imperative that Tesla either finds a way to release V13 on HW3 cars or offers an upgrade to AI4 so that the 6M cars can run V13. They can charge a small amount, say $250 for that upgrade but without that support, they will not be able to achieve large scale road safety for years until normal or even accelerated attrition eliminates those 6M older cars.
why is insurance for telsa so expensive? aren't they safer?
The most important part why FSD take rate is going to 100% is the car can go charging itself at power off-peak period.
People today pay more than $5k for passive safety systems in a car, like seatbelts, airbags, ABS, shatter proof glasses, traction control, crump zones, roll cages, collapsible steering. None does much to prevent an accident, just protects passengers life’s in case of accident. A single accident avoided by FSD exceeds its cost by large, not reven considering the lives saved. No doubt to me that one day will be mandatory, or at least an option in each car sold.
Why hasn’t insurance expanded?
Good question😢
Good question
Same reason it took Biden 4 years to approve the first lithium development in Nevada. And now, they're looking to block it longer because of a fictitious snail.
Question to the cyberbulls/cern:
To me there seems to be a pricing analogy between semis/busniss vehicles and cybercabs/robotaxi.
After all both are vehicles, that busniss-customers aim to earn cash with.
If the above is true: how are semi-sale prices formed, and could the same way be transfered to cybercabs?
We can simplify this. How many CEOs were hired to implement a smart car of the future that will include self driving? Current CEOs were hired to manage the existing business that they were given and this will be "all new business." Wouldn't the first task of an OEM be to think about hiring a new leader to reorganize the company into one capable of making a viable car to sell in this new future? This means working out new suppliers of hardware and of the software. A whole new set of licensing agreements. Realigning all new departments and the whole company. This feels like decades in the making. A brand new company might have a shot at succeeding while a long existing OEM will have to first be willing to remake themselves.
Tesla engineers would be wasting their time trying to teach OEM ICE companies how to integrate the FSD hardware and software suite. Would be like teaching physics to kindergartners. OEMs would basically be building Tesla clones from scratch with Tesla built hardware and have $10K or more added to the manufacturing cost. Easier for Tesla to build more Teslas than to sell hardware to OEMs. Apple briefly licensed Macintosh hardware to others long ago and it was a nightmare.
FSD in CA without safety score vs CO where safety score applies, only saved us $40/mo but could add $50 in a month with a forward collision warning (often false positives).
Cern Basher is amazing at explaining everything he creates (numbers, graphs, and statements with facts)
Omg, my mom is 65 yeas old this year, and she has nerves driven a car in her life simply because she thinks driving could be risky and if something happens, who’s gonna take care of me and my brother? Thus she gave up the thought of owning a car and learning how to drive. Now I am an adult and a long-term TSLA investor, I can’t wait to buy her a TESLA with FSD or even a Robotaxi so she doesn’t have to take bus or walk, her life experience will be changed profoundly, I am so touched by knowing this is going to happen. Thank you guys and Elon😭
we talk a lot about OEM and FSD - what about the car just having only one internal software unlike ford saying 170 vendors parts have separate software.
I don´t know how they do it but for the last 2 years, if you are using satnav to get to your destination, the system can access traffic speed and issues and advise a change of route to avoid the delay on your current route.
None of the legacy automakers have been able to deploy OTA updates and probably never will. For them to develop the hardware for FSD is going to be an impossible task unless Tesla gets involved and sells them the hardware as well. I don't think it is worth it for Tesla to do this hand holding. The big 2 are so used to being bailed out by the government, so they would rather do that. Tesla should go its own way and deploy cybercabs around the world.
I am thinking that the conclusions here were a bit US centric, and missing what could happen in other countries.
- FSD may take some time to train for many countries, and each will have regulators that will need to be convinced that it will work on their roads, ditto for Cybercab.
- Japan, Korea, Germany, Italy and India may well decide that they need, for strategic reasons to retain control of some auto manufacturing capability, as it feeds directly into defense capability.
- EV's not currently useful for people who want to tow large caravans, boats, horses or trailers over long distances, and perhaps the differing form factors of the towed things could make FSD very difficult.
- Cybertruck will be far less popular in many places - Japan, much of Europe and the UK for starters - it is simply too big for the roads and parking spaces there.
If Tesla manages to significantly lower their production costs on some models, they could keep the $5K Margin on hardware and include FSD for an extra $5K, preventing buying their cars without FSD. Then a single limitation like the impossibility to drive more than a mile without any passenger inside the car could allow Tesla to ask for a subscription if the owner wants to use their vehicle as a robotaxi (or send it far away to pick someone up). That way the buyers get a cheaper car and have FSD by default to keep them and others around safe, Tesla sells FSD with each car and this can apply to other OEMs using FSD.
I think Honda and Nissan are merging in order to get an FSD deal.
No one talks about additional safety when most cars are autonomous due to fact they will able to “talk” to each other
Lots of people discuss this.
Tesla should label it as a "safety system"
IN REGARDS TO THE OTHER GUY QUESTIONING USING OTHER OEMS AT 45:00 IN THE VIDEO. The other problem is the other automakers can not make evs profitably and the question will they ever be able to, is it worth trying to rely on them if there is a chance they will go bankrupt anyway before they can get their robocabs into the market.
We should also explore how OEM’s can duplicate Tesla data lead
They can’t duplicate the billions of miles of training. They will have to choose Tesla in the buy vs build choice.
“When you turn a depreciating machine into a money making machine” - Cern Basher
I think they decided to keep the FSD price lower for personal use, and get the value from robotaxi in the revenue share.
All it will take to break the dam is one successful licencing deal.
I believe Legacy Auto is interested. However as mentioned, it will take a few years to incorporate the cameras and inference chips into their cars to licence FSD, in other words... its very expensive.
So they'll likely wait until Tesla has proven the concept by successfully launching unsupervised in a couple of cities. Even if it means paying more by not being the first one to sign on the dotted line, they feel its prudent to wait for confirmation before committing what could potentially be a very large investment.
Once proof of concept is in place, they'll flood in. They simply cannot allow Tesla to be the only car manufacturer to have unsupervised FSD because the gap between haves and have nots will simply be too great
Ok thanks
All Tesla needs is for one of manufacturers to sign on the dotted line for fsd!
Don’t you think CyberCab fleets will be so hyperlocal that the fleet operators are more comparable to fast food franchises than car rental companies? Tesla themselves will be operating the largest fleet.
If there is a need for Mega Packs, I would go ahead to produce more Battery making plants. This can not only benefit Tesla but the consumers to reduce their energy cost plus the environmental solution to cleaner air in conjunction with solar and wind power
Maybe Tesla can license FSD on a quota basis that won't hurt the Tesla owners that wish to be involved in Robotaxi. Why hurt The Tesla owners that have bought a 2nd car just for that purpose. Thanks to Cern and Herbert
To be clear, I don’t think FSD would solve OTA software updates for other OEMs unless they buy all of their electronic components from Tesla.
Yes, it would help. But, the real issue is that the firmware on all the various components are built on separate codebases, separate processors, and code languages created by their suppliers. And they have no immediate control over the way the code works in each of these devices. They don’t even have the source code. These devices are just “black boxes”. It’s a big mess.
Tesla FSD would simply not be compatible with those "black boxes". OEMs would basically need to build a Tesla clone with Tesla produced hardware and wiring harness. Not very profitable when Tesla is a competitor.
There will some unemployment assistance tax for each displaced taxi driver if FSD is used in my state.
Sounds socialist 👺 to me ! 🤮
What happened to the road apple shovellers in New York?
I have long thought that Tesla will stop making cars for sale, focus completely on Cybercab, drop any idea of licensing FSD and make a fortune for shareholders given the potential profit margins. It's a no brainer and completely meets their stated mission of sustainable energy
Is there a clip of Jeff's insurance savings story, because my mom has a H4 MY but doesn't trust fsd yet and doesn't want to pay for it.
Shs should be able to try it for free soon enough if not already.
Consumers will initially just want to buy vehicles which have the CAPABILITY to run Tesla FSD, with all cameras, etc. The Consumer might not activate the FSD yet, but eventually they will.. So that's it's own potential profit situation than doing an "everyone gets FSD" model. Lastly, what's the cost that these other car manufacturers who are adopting Tesla-inside FSD systems will be paying to license Tesla tech?
Go Herbert!! Thank you!! 😎👍
I am sure there are others like me that have a Tesla, drive 5-10K miles per year, and do not see a great cost advantage of FSD because 60-80% of my driving is long trips (500 miles each way). Cruise control, which is already included with purchase, does a great job on long freeway drives without the $100/month or FSD price. In addition, if I purchase fsd outright and have a significant accident, there goes my FSD cost. Please give me a constructive argument against my logic. I am retired so daily driving is minimal.
Agreed
No matter what the OEMs do with FSD they have no match for Tesla’s factories. Would it be possible for Tesla to license factory design? 🤔
Tesla should 100% NOT share FSD. One big risk is giving it a bad name and bad press through sub-par implementation. Just so other car companies can profit/survive? 👎
Tesla really needs to sell the entire sled to the other car companies. This would allow them to have full self drive fully Incorporated, as well as OTA. Then, all the other car company would need to do is create the body shell to put over the sled.
how many cybertruck pre order two million ? now where the sale now?
Scenario one.
The ridiculous cost of insurance now day's (thanks for my good driver discount Geico! ugh) will make Robotaxi a more viable alternative to owning a car.
What if Tesla works closely with Magna to produce CyberCabs and Optimus.
IMO "the Cyber-Cab" revolution will eventually eliminate MOST desire to own cars and all the costs 👺 and liability of car ownership .... And, worrying about theft and dents, and so on and so on...
You want to get a cab every time you want to duck out to the store, or head to a friend’s place?
@@oggyoggy1299 If it costs almost NOTHING... And it takes me door to door... And I dont have to worry about demon-crats breaking into my car while I am inside the store... and I dont have to worry about accident liability... And I can focus on other things on the way there and back... And so on and so on... The answer is YES 😊
I don’t think licensing will happen with any American auto makers. It’s too much for them to install the chips, eight cameras, etc.. The only one might be BYD.
Tesla will be the Microsoft OS for automobiles
Aren't Ford working on a new EV mass production ready platform in a skunk works headed up by an ex-tesla employee. What if this new ford car has ground up been designed for, and has FSD hardware included already so, FSD ready when it is launched?
I generally agree with CERN, but I would note that no company lasts forever, even Tesla. I know the Ford and Porsche families would disagree, but I think they are wrong. Eventually, a new or better idea, and/or better management will come along. I acknowledge that it may be many years from now, but we should not assume that it will never happen.
Tesla producing 10 million vehicles is going to take a long time. Factories have to be built. It takes time to find sites, get approvals, build the factory, and then ramp it. AND right now there are no new factories on the drawing board. Until demand goes up, Elon is not building anymore factories.
maybe a business could be developed to retrofit any existing car to FSD from Tesla
Its. "CERN💪❤ Time" ❗ 💃🕺
How many systems don't use Windows architecture? This will be similar to Microsoft integration in the 90's.
Guys PLEASE Consider this Productisation ….. 2 very different Autonomy s/w packages .
1) ….FSD - as we know it today - at today’s prices or lower …..FSDS (Supervised ) demands seat pressure on drivers side to work : the best ADAS.
2) FSDU - (Unsupervised) no seat pressure demand - allowed to enter Tesla Robotaxi fleet & make money for owner . Different Price but must be related to $ yield :Yield depends upon Miles Driven & therefore must be priced PER MILE Driven using the s/w.
Nb Cern’s model misses the Margin that exists on EVERY sale in the channel . (EVERY DEALER WORLDWIDE) Ergo not a viable commercial comparison .
those CEOs whose bonuses are based only on a yearly basis, there is no incentive to think longer term. Licensing and implementing Self Driving will take over a year, it will be 5 years and longer. Plus a lot of effort... that no CEO wants to think about. They will prefer to have "the next guy" to worry about it.
I really doubt that other oem manufacturers could integrate the tesla hardware and software stack into their vehicles. They are really outdated compared to teslas computers on wheels
Adoption FSD can be difficult because especially German car manufacturers have the "not invented here syndrome". They want to invent it themselves and think they can do it better...
I lean toward the scenario where ICE OEMs are targeting selling POA EVs with fsd. But by the time they can bring that to market, CyberCabs have begun to erode POA sales. Tesla is disrupting their own EV sales business. Why bother trying to help OEMs sell into a market that is withering away?