According to sources, the first autopilot chips jointly developed by Tesla and AMD have entered the final testing phase and it is expected that full automatic driving will be realized by 2019. Although Tesla and AMD did not confirm or deny the cooperation rumors of the two sides, there are various signs that the introduction of AMD can well reduce the dependence on Nvidia, which is also very much in line with business logic.

At the same time, it was reported that Tesla will no longer use Nvidia’s chips to support its large-scale in-vehicle entertainment systems and instead use Intel’s chips. Although this news has not been confirmed, Tesla and Nvidia's "honeymoon period" does not seem very smooth. According to analysts, if Tesla is really "indifferent", it will cause potential losses of hundreds of millions of dollars next year.

In this regard, the recent Yellow Emperor in the field of AI chip (especially Nvidia founder and CEO Huang Renxun) do not know how to feel.

In fact, most companies have become accustomed to Tesla’s “passion for love”. In July of last year, after Mobileye and Tesla broke up, Nvidia "entered the battle" and reached a strategic partnership with Tesla. However, Tesla has his own careful thoughts, he does not want to rely on human breath, there will be a plot later AMD involvement.

As a giant in the automotive industry, Tesla has never stopped the layout of key areas. Even with the cooperation between Nvidia and AMD, Tesla has always insisted on independent research and development of AI chips. According to news, the main reason for Tesla’s cooperation with AMD is to value AMD’s intellectual property rights in AI chips. The rumors of the “break up” with Nvidia also reflect Tesla's desire to emphasize autonomy in the AI ​​chip field.

Nvidia's layout

Nvidia is popular as the "popular niche" of the AI ​​chip manufacturing industry. But driven by the prospect of automated driving today, many partners have secretly started and have their own calculations. And Nvidia is quite helpless. After all, even though it owns strong AI chip manufacturing technology, it has insufficient experience for the automobile manufacturing industry. Otherwise, Nvidia has already deepened into the field of automatic driving.

Of course, even if Tesla is distracted, Nvidia is still very strong in the field of general-purpose computing GPUs and has established cooperative relationships with a number of automakers.

From the market point of view, Nvidia had entered the list of cooperators of many car companies earlier with Tegra series processor technology. He did not lack cooperation partners, such as the Audi A8 who had been cooperating. This year, the latest Tegra K1 chip developed by Nvidia has been adopted to enhance the vehicle's ability to look around and handle the impact. Shortly afterwards, Nvidia also launched the DrivePX2 and SoCXavier based on the Volt architecture GPU. This enabled Nvidia to regain its position in the field of automatic driving. Even the global automotive companies Bosch and ZF have reached cooperation with Nvidia, and will use the Xavier platform to develop on-board computers. This will become a turning point in the development of autonomous driving.

Nvidia founder and CEO Huang Renxun (Source: Nvidia website)

Tesla's backhand

Of course, if Tesla said that he does not research and develop AI chips, I think most people are skeptical. This is based on Tesla's talent pool in the AI ​​chip field.

Among them, the most direct expression is the AI ​​chip leader who joined Tesla last year: Jim Keller. As head of Tesla's automated driving hardware and software, Keller has extensive experience and tough technology. In his early years, he worked for AMD and participated in the construction of the X86-64 architecture. He then joined Apple and was responsible for the design of the A4 and A5 processors.

In addition, Jim Keller is not alone in the battle. His team includes: Peter Bannon, Apple's top chip developer; David Glasco, AMD's server chip architecture senior director, has worked for Intel, IBM, and Nvidia; AMD cloud service architect ThaddeusFortenberry; AMDFellow, CPU chief architect DebjitDasSarma; AMD strategic cooperation vice president KeithWitek; AMD machine learning chief engineer JunliGu.

Can you imagine that such a luxurious chip team configuration, you said Tesla did not own the idea of ​​chip development, anyway, I do not believe it.

Actually, Tesla’s situation has been unfavorable, especially since last year a driver who used Tesla’s Autopilot function died of a car accident and once pushed Tesla onto the cusp, even if the final investigation showed that the driver was making mistakes but it was safe. Sexuality has also become a major point of view for consumers. Tesla's eagerness to find partners is due in large part to the invisible pressure of opponents and the need for new self-driving vehicles to change the public’s image. Such as Google's Waymo, to know that Google's R & D strength in the AI ​​for all to see, its main push TPU has now developed to the second generation, Google AI research and development of the main direction is to solve the problem, it has a unique technology in hardware and mobile software . In addition, auto-driving R&D companies such as Apple, Uber, and Baidu are all eyeing.

Autopilot has become an important market for major car companies to compete in the future. Tesla is no exception. For the core data processing and security assurance issues, Tesla naturally does not want long-term dependence on the outside. Tesla’s own chips are probably on the road.

Insulated Terminals

Insulated Terminals,Terminals,High-quality insulated terminals

Taixing Longyi Terminals Co.,Ltd. , https://www.longyicopperlugs.com