{{ Chinese officials told Tesla that Beijing has tentatively approved the company’s plan to launch its “Full Self-Driving,” or FSD, software feature in the country, people familiar with the matter said.
The U.S. electric-vehicle maker will deploy its autonomous driving services based on mapping and navigation functions provided by Chinese technology giant Baidu, the people said.
The partnership clears an important regulatory hurdle for Tesla to offer its driver-assistance system there. Partnering with a Chinese company helps ease regulators’ concerns over any data-security risks, the people said.
Bloomberg earlier reported the deal with Baidu, which in addition to its core search-engine business has expanded into autonomous driving and artificial intelligence. }}
Similar to other US business strategies. If regulation makes your product costly, use other countries with less regulation to produce your product.
In this case, it’s to develop and perfect, but the principle is similar. China absorbs the initial rollout, eventually we’ll bring the technology back when regulations are relaxed.
Assuming there are three markets that will decide the fate of autonomous driving, China, EU, and the USA, my question would be “Which bureaucracy is less likely to obstruct progress?” My guess, “China.” During the earnings conference call, Elon Musk said that Tesla was in talks with one or more automakers to license FSD. The current news points to BYD as the primary candidate. The two biggest EV makers can set the standard for all others. Like Tesla set the NACS charging standard in the US.
The H100 is not used in the cars. Only in training the neural networks. Other chips can do that even if less efficiently, maybe Dojo as a service. (DaaS).
I thought they were using the H100 for inferencing also? But you are right they probably can do it with a less powerful chip and it wouldn’t make that much difference. But I think Musk was saying that each country has it’s own rules so Training will have to be done in each country. Dojo as a service is interesting, do you think China would allow the data to be taken out of the country?
Yowza! This is a pretty big deal if it comes to pass and it explains much of what has been going on with Tesla lately. Negotiations with the Chinese government and Baidu must have been going on for awhile. As a deal became more likely, I can understand why Tesla would make a rapid shift in its strategic plan to put FSD/robotaxi ahead of the $25K vehicle. Tesla potentially now has the opportunity to dominate the world’s largest auto market with what is essentially a software upgrade.
It looks like Tesla is going to launch FSD in China in the next year for its existing cars. That is potentially a lot of high margin revenue in software sales. It will develop this into an autonomous driving service under the auspices of the Chinese government, creating more potential high margin revenue while also gaining the experience/data to satisfy regulators in other countries.
In the meantime, it will be developing and bringing into market lower price versions of existing models to maintain sales in the west as it initiates and ramps production of its next-gen robotaxi. With the latter, it can choose to follow the Apple model and create vehicles optimized to its software, or it can deemphasize car production by licensing the software and letting the partnered OEMs race to the bottom making commodity hardware. An Apple or Microsoft future.
Still just rumors and hints, but if Tesla can enter China with FSD and robotaxis in this period of intense competition and Western sanctions, how is Musk not worth $50+B?
If he would do the job for $10B, then he’s not worth paying $50B. And he’d do the job for $10B. So just pay him $10B. Or something less, since he’d probably do the job for less.
Because he did not earn it–yet. Wait to see IF the China self-driving test really provides safer driving. Tie a moderate bonus to THAT. Then look for introduction of self-driving in other countries–and tie bonuses to successful outcomes in those other countries.
FSD and robotaxis are two different (but obviously related) things.
FSD (supervised or beta) currently requires a hands on driver…this is what this deal is announcing, AFAIK.
Robotaxis, as everyone knows, requires no driver (level 4 or 5) plus lots more stuff (an app to summon and pay, charging and parking lots, cleaning, emergency crews to rescue stalled, blocked or damaged cars).
So, this is a required milestone on the way to a robotaxi, not yet there though.
I’m not sure you understand just how much money that is, how it relates to the history of executive compensation, and how it relates to life-time profits of Tesla overall. In short, its an absurd amount of money to compensate any one individual for.
Tesla’s stock was up $26/share today to $194. Investors seem to be judging this as a big positive. Still, there are big bureaucratic roadblocks to FSD approval in the US.