Cao Yuan is from the Vice-Pilot Temple
智能车参考|公众号AI4Auto
At the tail of the Beijing Auto Show, Musk landed in Beijing.
After 11 months, Musk arrived at Beijing Capital Airport by private jet. Subsequently, the pick-up and meeting pictures were exposed.
Coincidentally, on the same day of Musk's visit to China, Tesla's announcement on passing the data security test was released:
The domestically produced Model 3 and Y all meet the requirements for automotive data security compliance. Tesla, it's time to lift the ban.
Moreover, Musk is not only busy visiting China, but also announced a major investment plan for Tesla this year: 10 billion US dollars, about 72.4 billion yuan, all for training automotive AI.
Behind this is the ultimate goal: autonomous driving.
Tesla lifts the seal
Shortly after the news of Musk's visit to China broke, his private jet landed in Beijing.
According to the news, Musk's visit to China was invited by the China Council for the Promotion of International Trade, and Musk and Ren Hongbin, president of the China Council for the Promotion of International Trade, discussed the development of Tesla in the Chinese market and opportunities for future cooperation.
In addition, Premier Li Qiang also met with Musk at the Diaoyutai State Guesthouse, and the specific content of the conversation was not exposed, but Tesla and Musk expressed their honor and gratitude through their official accounts.
On the same day, the China Association of Automobile Manufacturers and the National Computer Network Emergency Technology Processing and Coordination Center jointly issued an announcement about Tesla, titled "Notice on the Testing of Four Safety Requirements for Automotive Data Processing".
In accordance with the relevant provisions of relevant regulations and standards, the two institutions have tested the intelligent networked vehicles listed in China from 2022 to 2023, and the purpose of this notification is to release the monitoring results.
The results show that 19 models of domestically produced Tesla, namely the 2022 and 2023 versions of Model 3 and Model Y, all meet the four safety requirements for data processing, namely:
- anonymization of face information outside the vehicle;
- Cockpit data is not collected by default;
- in-vehicle processing of cockpit data;
- Conspicuous notice of the processing of personal information.
In other words, Tesla complies with domestic regulations on the security and privacy protection of automotive data processing, which will help lift the restrictions on Tesla's driving and parking in airports and other places.
Subsequently, Tesla said that restrictions on intelligent networked vehicles such as Tesla have been lifted one after another.
Looking back at the list, Tesla has also become the only compliant foreign-funded car company, and the rest of the car companies also include BYD, Ideal, Lotus, Hezhong New Energy and NIO 5 companies, with a total of 57 models.
However, it should be noted that the results are based on the principle of voluntary inspection by enterprises, and it is not clear how many car companies have submitted for inspection this time.
For Tesla, data security compliance not only means the lifting of the ban, but more importantly, it is related to the landing of FSD, and there are no obstacles.
This is also closely related to Musk's and Tesla's goals for the next stage.
Elon Musk: Invest $10 billion in automotive AI this year
A well-known Twitter technology blogger recently revealed that Tesla invested $1 billion (about 7.2 billion yuan) in AI training computing in the first quarter and doubled its computing power.
Just when netizens began to be surprised by Tesla's big spending, Musk forwarded the news and said that $1 billion is just sprinkling, and Tesla's planned investment in AI training and inference this year is actually $10 billion (about 72.4 billion yuan).
In addition, the training of reasoning ability is mainly for automobiles, that is, Tesla's self-driving ability.
It seems that what Musk said before is true, and the top priority of Tesla's development this year is autonomous driving.
And with so many years of accumulation, Tesla has been fully prepared.
As we all know, Tesla's route to achieve autonomous driving is iterative, iterating from L2 to the top.
Tesla has reserves for the systems, data, computing power, etc. required for iteration.
The latest version of FSD V12.3 has been able to achieve intelligent driving in three domains, which is the highest capability in the L2 level, which not only includes the basic algorithm system, but also means that a steady stream of data will be collected.
As of April 6 local time, the mileage of FSD has exceeded 1 billion miles (about 1.6 billion kilometers).
The inflection point of this rapid growth is the time when FSD Beta is suddenly pushed on a large scale. Now that the full capability is live, it's likely that the use of miles will usher in new growth.
And, having large-scale data is one thing, but more important is training, or the deployment of computing power.
Tesla is prepared for this.
One is Nvidia H100, Musk revealed on the Q1 conference call that the company has about 35,000 pieces, and the other is Tesla's self-developed Dojo supercomputing system, which was put into production in July last year, and it is said that a chip can provide 362 TFLOPs of computing power, and the latest physical picture is also unveiled in the financial report.
In addition, last year, Tesla announced that it would build "the first of its kind" data center and open up positions such as senior engineering project manager for data centers. It was later revealed that the data center, located at the headquarters in Austin, Texas, was built specifically for Dojo.
Therefore, although $10 billion may sound like a lot at first glance, as Musk said, if you invest in training computing resources, build a large number of data transmission channels, and store massive videos to achieve universal autonomous driving, 10 billion may not be enough.
However, Tesla has the money, resources, and can afford it. Musk believes that this investment is nothing compared to the value of Tesla cars on the road with a total value of about 25 trillion US dollars (about 181 trillion yuan).
And although Musk said that it was not enough to spend, if you look at it in the industry, Musk still feels that Tesla's level is unbeatable.
Any company that can't invest that much, or can't invest efficiently, can't compete with Tesla in this space.
For example, next door to Apple, which spent tens of billions of dollars in ten years to build a car, but in the end there was not even a shadow of a car, let alone the realization of autonomous driving (doge).
Therefore, Tesla basically has the money, resources, and capabilities needed to realize autonomous driving, and will officially start a sprint to realize autonomous driving this year.
In order to better collect more data, it is almost inevitable that FSD will enter China and lift the seal.
Just now, He Xiaopeng, CEO of Xiaopeng Motors, the top echelon of intelligent driving in China, expressed his views on Tesla's FSD entry into China on Weibo.
He Xiaopeng said he was very welcome. He believes that whether it is FSD, Xpeng's NGP, or NOA of friends, they are all representatives of new quality productivity and should support the blooming of a hundred flowers.
So among these flowers, which one do you prefer?