Elon Musk admits other automakers don't want Tesla's over-hyped FSD software

[DragonSlayer101]

Facepalm: Elon Musk has finally admitted what auto industry insiders have said for years: no other automaker wants to license Tesla's Full Self-Driving technology. The admission comes after years of bragging by Musk and other senior Tesla executives about how great the software is and how major car companies are supposedly lining up to use it.

This week, Musk admitted on X that since offering to license Tesla's self-driving technology for a small fee, there have been no takers. Most rival automakers want nothing to do with it, and the few that did show interest wanted to run small pilots with no immediate plans to bring it to scale.

Musk recalled previously warning automakers they could face an existential crisis without FSD, but most still turned him down. One major automaker reportedly wanted to implement FSD in an extremely limited trial with "unworkable" conditions for Tesla, ensuring the talks went nowhere.

I've tried to warn them and even offered to license Tesla FSD, but they don't want it! Crazy …

When legacy auto does occasionally reach out, they tepidly discuss implementing FSD for a tiny program in 5 years with unworkable requirements for Tesla, so pointless.

– Elon Musk (@elonmusk) November 24, 2025

Musk stopped short of naming the automaker, but reports from last year claim Tesla was negotiating with Ford over an FSD licensing contract. In June, Ford CEO Jim Farley confirmed those talks, telling journalist Walter Isaacson that the deal fell through after the company evaluated Tesla's and Waymo's driver-assistance software and found Waymo's LiDAR-based system to be far more advanced than Tesla's, which relies solely on cameras.

Tesla charges a one-time fee of around $10,000 for its Full Self-Driving (FSD) suite, and most customers gladly pay for what is legally just Level 2 driver-assistance software. The technology allows cars to steer, change lanes, park, and respond to traffic lights autonomously, but it still requires human supervision.

Despite Musk's optimism, FSD and its predecessor, Autopilot, have often proven less than reliable. The technology is currently under investigation by the National Highway Traffic Safety Administration following reports of vehicles in FSD mode running red lights and driving the wrong way on highways and suburban roads.

In August, a judge found Tesla liable for a death linked to Autopilot and ordered the carmaker to pay $242 million in damages. The company is also under investigation for four other crashes reportedly involving FSD, with the probe covering up to 2.4 million vehicles.

Permalink to story:

Elon Musk admits other automakers don't want Tesla's over-hyped FSD software

 

[yRaz]

This is a 2 sided problem. 1) it just isn't good. 2, and most importantly, remember when Netflix came out and everyone created their own streaming service? Everyone wants to create their own FSD drive tech so they can charge for it instead of having to license someone else's tech. Very little of it have anything to do with Elon Musk or Tesla directly.

 

[brucek]

If I was an auto-maker, and I could re-sell someone else's tech at a profit, while letting them absorb all the initial high cost of R&D and the liability of mistakes, I'd jump at the chance to do that vs. take on those risks & expenses on my own.

Sure, much later, when it was all just proven commodity stuff, bring it in house then. But today it's all cost all risk and not much upside unless you truly believe you're positioned to do it better by yourself than the rest of the combined industry can do it by combining R&D funds. Just doesn't make sense.

For a smaller stakes decision, look at map data - if you're a car brand you license easily available ready to go data, you don't reinvent the wheel yourself. (Just like you buy tires from companies who specialize in that.)

Anyway, my optimistic read of this article is that even insiders feel Waymo is making good progress and much better options than Tesla's "FSD* (*just kidding actually supervised assistance only)" BS may soon be coming to more cars near me.

 

[Axeia]

Tesla messed up when they removed LIDAR. It was 100% a cost cutting measure as having more data to work with is never a bad thing and now part of the problem of them not keeping up with Waymo.
I can totally see Alphabet [Google] (they're the ones making Waymo) providing cars with FSD-software like how they're already providing phone makers (and carmakers as well really) with Android at the moment.

Having one company already proficient in software fully commit to the task rather than dozens of carmakers trying to do so is likely to end up safer for the rest of us that have to be outdoors interacting with these cars. Knowing how other Waymo cars will predictably behave would make things easier as well.

I could also see other carmakers forming regional alliances to provide a counterweight so they don't end up with yet another heavy reliance on American software though.

 

[Theinsanegamer]

Perhaps if FSD ever lived up to the promises people would be more interested?

Although we're going the other direction, people are waking up to the nightmare that these systems are and actively want as little tech in their cars as possible. The constant BONG BONG BEEP BEEP FLASHING LIGHT WAHH WAHH NAG NAG of modern cars is getting really annoying.

It's also just an objectively terrible name. There is no full self driving, so long as you have to keep your eyes forward and hands on the wheel.

Tesla messed up when they removed LIDAR. It was 100% a cost cutting measure as having more data to work with is never a bad thing and now part of the problem of them not keeping up with Waymo.
I can totally see Alphabet [Google] (they're the ones making Waymo) providing cars with FSD-software like how they're already providing phone makers (and carmakers as well really) with Android at the moment.

Having one company already proficient in software fully commit to the task rather than dozens of carmakers trying to do so is likely to end up safer for the rest of us that have to be outdoors interacting with these cars. Knowing how other Waymo cars will predictably behave would make things easier as well.

I could also see other carmakers forming regional alliances to provide a counterweight so they don't end up with yet another heavy reliance on American software though.

They're not the only ones. There are problems with LIDAR, one is that its expensive, two is that re configuring it after accident repair is an actual nightmare, and three that the most effective implementations look goofy and fly right int he fact of aerodynamics.

I agree though that you need LIDAR for anything remotely close to actual FSD, since cameras are just as vulnerable to vision issues as human eyes. Arguably moreso.

 

[Megalomaniac]

I've used FSD in the cyberbeast and I've used it in the new Model Y.

It's very impressive, but it's overly ambitious - especially here in NYC.
FSD will put you in dangerous situations. On paper it's just an "intersection". In Real Life, everyone is speeding and you can't be sure if they can stop fast enough to not kill you.

I don't forsee driverless cars taking over any time soon.

For now: FSD is an advanced cruise control, but you still need to be behind the wheel.

It will be a long, long, long time before I can get in a driverless cab drunk and tell it to take me home.

 

[Plutoisaplanet]

Perhaps if FSD ever lived up to the promises people would be more interested?

Although we're going the other direction, people are waking up to the nightmare that these systems are and actively want as little tech in their cars as possible. The constant BONG BONG BEEP BEEP FLASHING LIGHT WAHH WAHH NAG NAG of modern cars is getting really annoying.

It's also just an objectively terrible name. There is no full self driving, so long as you have to keep your eyes forward and hands on the wheel.

Well you don’t need your hands on the steering wheel at all, you just need to pay attention. There are now thousands of videos online of people putting in a destination, tapping “Start Self-Driving”, and doing nothing until the car parks itself at the destination. Drivers aren’t touching any controls at all. It’s not perfect yet, but it certainly appears to be getting closer with each software update.

Here are just 3 examples of the first results on YouTube uploaded today (multiple FSD drives each, no human driving). I could find only one case where FSD told the driver to take over briefly (who just re-enabled FSD), and no other interventions in any of these:

 

[p51d007]

I'm too old school. Grew up in the 60's "muscle car" era, started driving in the 70's. YOU drive the car.
Heck, I still drive a manual transmission vehicle and even got my boss to give me a manual transmission car to drive for work! Heck, it's more of a "anti-theft" device now because who knows how to drive a stick these days LOL.
Other than cruise control once in a while, I'm just not into all of these fancy self drive, 3923 sensors/cameras nonsense.

 

[OortCloud]

Why is Teslas share price still so high? Unreliable and uninspiring cars, their one USP - electric now long gone, huge competition from China, a boss who outside the US is basically a pariah for his horrific political leanings and awful petulant behaviour, Self Driving that will never work, The Cybertruck..., it's a catalogue of issues.

 

[Theinsanegamer]

Why is Teslas share price still so high? Unreliable and uninspiring cars, their one USP - electric now long gone, huge competition from China, a boss who outside the US is basically a pariah for his horrific political leanings and awful petulant behaviour, Self Driving that will never work, The Cybertruck..., it's a catalogue of issues.

Because they are still the only profitable EV manufacturer in the West. BYD makes unreliable trash whose reputation is quickly sliding, and in the world of investment the crying of Redditors that everyone is a big bad Not-See holds no bearing.

Well you don’t need your hands on the steering wheel at all, you just need to pay attention. There are now thousands of videos online of people putting in a destination, tapping “Start Self-Driving”, and doing nothing until the car parks itself at the destination. Drivers aren’t touching any controls at all. It’s not perfect yet, but it certainly appears to be getting closer with each software update.

Here are just 3 examples of the first results on YouTube uploaded today (multiple FSD drives each, no human driving). I could find only one case where FSD told the driver to take over briefly (who just re-enabled FSD), and no other interventions in any of these:

So it watches your eyes instead of your hands, but it still does the same nagging everytime you look away.

A full self driving car is one I can get in, lay down, and take a nap with no need to ever intervene, no matter the weather or location. FSD is still getting into accidents and ramming into barricades, and still wont work in snow or if for any reason there is a camera issue. It's a fancy assistant, but its not proper FSD.

 

[sreams]

Tesla messed up when they removed LIDAR. It was 100% a cost cutting measure as having more data to work with is never a bad thing and now part of the problem of them not keeping up with Waymo.

No Tesla ever had LIDAR, so how was it removed?

 

[Plutoisaplanet]

So it watches your eyes instead of your hands, but it still does the same nagging everytime you look away.

A full self driving car is one I can get in, lay down, and take a nap with no need to ever intervene, no matter the weather or location. FSD is still getting into accidents and ramming into barricades, and still wont work in snow or if for any reason there is a camera issue. It's a fancy assistant, but its not proper FSD.

For now. The big safety limitation when FSD causes a collision appears to be limited to very low speeds where precision is needed while navigating around objects. FSD could very easily become a level 3 driving assistant where the vast majority of the time you don’t need to pay any attention until it asks you to monitor while parking. I’d guess that happens in the next year sometime. But for now, it needs you to be looking at the road and it will nag you if you look at elsewhere in the cabin for at least ~7 seconds.

There may be situations when older versions of FSD are ramming into barricades, but I don’t think that happens on any current version. I’d be interested to see if this happens at all with V14. It’s still new, so I’d guess there aren’t any reports yet.

 

[Nobina]

It's unnecessary and promotes carelessness. I will drive myself, thank you very much.

 

[Puiu]

Tesla's FSD is broken and unsafe. Nobody wants a system that requires an expensive computer, buggy software and has no LIDAR support.

 

[DrSuess]

No they want FSD tech that works and can be delivered on schedule.

 

[DrSuess]

It's unnecessary and promotes carelessness. I will drive myself, thank you very much.

The fact they refuse to use LiDAR already makes it less capable and reliable than other systems that can operate in more adverse weather conditions. Tesla's RoboTaxis aren't performing as well as Waymo because they are using their own FSD tech. Tesla's RoboTaxis still have a human driver or a safety monitor in the front seat to meet regulatory requirements and supervise the system. Human drivers were supposed to be gone in June/July 2025.

 

[Melkor Unlimited]

FSD will put you in dangerous situations ...

Anything that does have control over car can put you into dangerous situations.

Line control assistant (or whatever name the car producer uses) trying to "correct" trajectory of the car when you go "too close" to the edge of the road because there is only so much space on a narrow country road and big semi is approaching on the other lane .. is enough to cause car crash.

 

[sreams]

Tesla's FSD is broken and unsafe. Nobody wants a system that requires an expensive computer, buggy software and has no LIDAR support.

All of these systems require an expensive computer. It doesn't matter how much data your sensors collect if the car doesn't know what to do with that data.

 

[Puiu]

All of these systems require an expensive computer. It doesn't matter how much data your sensors collect if the car doesn't know what to do with that data.

How many versions did of the hardware released and how many times did they promise that "this is the one!"?

 

[sreams]

How many versions did of the hardware released and how many times did they promise that "this is the one!"?

Your response doesn't even remotely relate to what I said.

 

[Vanderlinde]

I'm too old school. Grew up in the 60's "muscle car" era, started driving in the 70's. YOU drive the car.
Heck, I still drive a manual transmission vehicle and even got my boss to give me a manual transmission car to drive for work! Heck, it's more of a "anti-theft" device now because who knows how to drive a stick these days LOL.
Other than cruise control once in a while, I'm just not into all of these fancy self drive, 3923 sensors/cameras nonsense.

I have a quite modern A8 and I'm surprised by the luxury, comfort and all sorts of bells the thing has. It really makes driving a pleasure and with a lot of safety features your kind of safe in such a large tank.

Not saying I don't favor manual shifting and revving a V8; but the Automatic gearbox does a far better job consistent then I would do ; even at trips over over 5000km in a week.

 

[Puiu]

Your response doesn't even remotely relate to what I said.

It was related as to why the Tesla one is expensive.

 

[sreams]

It was related as to why the Tesla one is expensive.

You'd need to show that the Waymo computer isn't equally expensive for that to be a valid point.

 

[SRB]

On the subject, none of this can adequately compensate for rain, snow, etc. The human brain can use you eyes and put together images that have never been seen before. Training can't adequately cover it, and lidar and radar are blind in such conditions. AI to the rescue?????

P.S. Good to see the forums cleaned up a bit. Thank ;you.

 

[Puiu]

You'd need to show that the Waymo computer isn't equally expensive for that to be a valid point.

For Waymo you pay for the massive array sensors not the computer.

Waymo has 5 lidars, 29 cameras and 6 radars. A tesla has just 8-9 cameras. You can understand why Waymo is charging a lot (with a single Lidar being about 1000$), but you cannot make the same argument for a tesla.

Without Lidar or at least Radar sensors, Tesla will never reach level 5. It's a pipe dream that that they sell to gullible people.

And Waymo released many white/research papers that others can study and use.