Morgan Stanley: Dojo is Tesla's AWS, and the stock price could rise another 60%!

Wallstreetcn
2023.09.11 08:36
portai
I'm PortAI, I can summarize articles.

Morgan Stanley has upgraded Tesla's rating from "hold" to "overweight" and raised its target price for the next year to $400 per share, which is the highest price given by Wall Street investment banks. This represents a 61% upside from the closing price of $248.5 on September 8th.

Goldman Sachs bluntly stated that benefiting from the support of Dojo supercomputer for Tesla's autonomous driving business, Tesla's stock price will soar. The value of Dojo to Tesla is like AWS to Amazon.

Goldman Sachs upgraded Tesla's rating from "hold" to "overweight" and raised the target price for the next year to $400 per share, which is the highest price given by Wall Street investment banks. It has a 61% upside potential from the closing price of $248.5 per share on September 8th.

Recently, a team led by Morgan Stanley analysts Adam Jonas and Daniela M Haigian pointed out in a report that Dojo supercomputer can open up "new potential markets" for Tesla, allowing Tesla's market value to increase by over $500 billion. They wrote:

Dojo supercomputer is designed to process a large amount of data in training autonomous driving systems. It may bring "asymmetric advantages" to Tesla in this autonomous driving field with a potential value of $10 trillion.

From now on, software and services become the biggest value drivers for Tesla.

Goldman Sachs analysts believe that the currently highly anticipated areas include the next-generation fully autonomous driving system expected to be launched by Tesla at the end of the year, as well as the AI Day scheduled for early 2024.

Wallstreetcn previously reported that in 2019, Musk announced the independent development of a supercomputer for autonomous driving data training, named Dojo, using video data from Tesla's fleet for video training.

At the AI Day in 2021, Tesla unveiled its own supercomputer, Dojo, for the first time. Although Tesla already had a supercomputer based on NVIDIA GPUs, Musk emphasized that the development of Dojo was based on Tesla's self-developed chips.

Nearly two years have passed, and in July of this year, the Twitter (now X) technology disclosure account Whole Mars Catalog revealed that Dojo has officially started working. This news was also confirmed by Musk himself with a like.

According to Tesla's information, Dojo integrates 120 training modules, has 3,000 D1 chips, over 1 million training nodes, and computing power of 1 EFLOP, which is equivalent to 10^18 floating-point operations per second.

After Tesla revealed the production of the Dojo supercomputer in its earnings report, Musk also stated in a conference call that he expects to invest over $1 billion in this project by the end of 2024. Morgan Stanley analysts have expressed that the more research they conduct on Dojo, the more they realize that Tesla's value may be underestimated. They have raised their target price for the next year to a market high of $400 per share. According to data tracked by Bloomberg, the average target price for Tesla among analysts is $268.42.