Tesla Inc.’s Dojo supercomputer may add as much as $500 billion to the company’s market value through faster adoption of robotaxis and network services, according to Morgan Stanley.

Dojo can open up “new addressable markets,” just like AWS did for Amazon.com Inc., analysts including Adam Jonas and Daniela M Haigian wrote in a note, upgrading the stock to overweight from equal-weight and raising its 12-month price target to a Street-high $400 per share from $250. Tesla has already more than doubled this year and closed at $248.50 on Friday.

The supercomputer, designed to handle massive amounts of data in training driving systems, may put Tesla at “an asymmetric advantage” in a market potentially worth $10 trillion, they wrote, and could make software and services the biggest value driver for Tesla from here onward. The next version of Tesla’s full self-driving system, expected by year-end, and the company’s Artificial Intelligence day in early 2024 are worth watching, they added.

Tesla has been mentioning how Dojo gives it an edge in AI and self-driving technology since at least 2021. In July this year, CEO Elon Musk told investors that the carmaker plans to invest more than $1 billion on the project by the end of 2024.

That base-case target from Morgan Stanley would put the stock near its record close of $409.97 in November 2021. That makes the firm a notable outlier: The average price target among analysts tracked by Bloomberg is $268.42.

“The more we looked at Dojo, the more we realized the potential for underappreciated value in the stock,” the Morgan Stanley analysts said.