Tesla Invests $1 Billion in Dojo Supercomputers – Less Dependence on NVIDIA

Tesla is also still chasing records. Now Elon Musk wants to use the Dojo supercomputer to equip Tesla vehicles with their own infrastructure and thus leave the competition in the field of artificial intelligence behind.

• Tesla continues to focus on AI development
• Dojo supercomputers soon in top five worldwide
• Close cooperation between xAI and Tesla

In the past quarter, electric car manufacturer Tesla once again set delivery and sales records. However, the price cuts with which Tesla had driven its sales records were reflected in the development of profits. In the conference call with analysts and investors stressed Elon Musk once again that he would accept setbacks in profitability in order to keep cars affordable and to promote sales in times of high inflation and rising interest rates. The focus continues to be on reducing production costs and increasing production efficiency. The Tesla CEO also went into the “Full Self Driving” system, which is not considered an “autonomous driving system” in the industry, but as an assistance system. He also stuck to his version of autonomous driving, despite being overly optimistic in the past, saying, “I think by the end of the year we’ll be better than humans.” He also went into the often announced Robotaxi in “revolutionary design”.

Start of production for the Dojo supercomputer

With the production of its own supercomputer, which is scheduled to begin in July, Tesla wants to make itself technically independent of NVIDIA, whose A100 and H100 GPU chips have so far formed the basis for the Tesla systems. As IT TIMES writes, Tesla will replace the previously used GPU-based supercomputer from NVIDIA with the new, custom-made Dojo supercomputer with a chip architecture developed purely by Tesla.

The supercomputer for machine learning, which was already presented at AI Day 2021, is expected to be among the top five in the world by early 2024. As reported by Electrek and Teslarati, Dojo uses Tesla’s own chips and infrastructure as well as video material from the “Full Self-Driving Beta” version for processing huge amounts of data and developing the software for autonomous driving. According to the experts, Tesla could reduce its dependence on chip manufacturers with the inference chips and the AI ​​training based on the video data and take a big step in AI development.

According to a report by Bloomberg, the tech billionaire said he plans to invest over $1 billion in his supercomputer project by the end of next year. Although Tesla CFO Zachary Kirkhorn also emphasized that the investments were in line with the three-year outlook, Tesla shares fell significantly in view of the enormous sum.

xAI: Tesla, Twitter and the “true nature of the universe”

As CNBC reports, Elon Musk wants to advance the topic of artificial intelligence on both the “silicon front” and the “AI software front” with his new AI startup xAI. In both areas, cooperation with Tesla is an integral part. Announcing during a live audio session via Twitter on July 13, Elon Musk emphasized xAI’s goal is to “understand the true nature of the universe” — namely, the physical world and not just the internet.

The CEO of Tesla, Twitter, The Boring Company and xAI announced that with xAI he wants to emulate Alphabet’s DeepMind computer program AlphaZero, using text, image and video data from the social media platform Twitter.

Editorial office finanzen.net

Selected leveraged products on Alphabet C (ex Google)

With knock-outs, speculative investors can participate disproportionately in price movements. Simply select the desired leverage and we will show you suitable open-end products on Alphabet C (ex Google)

ttn-28