Analyst cheers the potential in the company’s Dojo supercomputer
Shares of Tesla Inc. climbed in premarket trading on Monday after Morgan Stanley cheered an artificial-intelligence opportunity for the electric-vehicle company and became the most bullish analyst on its stock.
Morgan Stanley analyst Adam Jonas upgraded Tesla’s stock TSLA, +10.09% to overweight from equal weight in his latest note, basing most of that newfound optimism on Tesla’s new machine-learning supercomputer, Dojo. He also lifted his price target to $400 from $250 per share, with the new target now the highest on Wall Street, according to a FactSet survey of 44 analysts.
Shares of Tesla climbed 5.7% to $262.37 in early trading, which puts the stock on track for its highest close since late July should that level hold.
The fresh recommendation was included in a larger note from a group of Morgan Stanley analysts led by Edward Stanley, who discussed the value of “finding plausible, scalable Moonshots that have not yet been fully discounted by the market.” They said Tesla’s Dojo project “could tick all the boxes.”
“Dojo specializes in processing visual data and is a tailored tool to solve autonomous operations, be that for vehicles, factories or humanoid robots. This adds Tesla to a list of companies ( META, +3.25%, AMZN, +3.52%, MSFT, +1.10% ) that have also developed their own, proprietary computing solutions,” Jonas and the Morgan Stanley team said in a note.
Tesla first unveiled “Project Dojo” in 2021 to train models for its self-driving efforts more quickly. Teaching a computer to operate a car relies on copious amounts of data, which is heavy on capital expenditure and is energy intensive.
“By designing their own computer chip (i.e. ‘custom silicon’), integrating it with purpose-built hardware and software, Tesla estimates that they can get an [approximately] 6x cost improvement per unit of computing power due to the much more efficient system,” the Morgan Stanley team said.
What makes Tesla’s Dojo system a “moonshot,” in their view, is the fact that “true innovation and closing the gap to [AI chip giant] Nvidia AR:NVDA is a remarkable claim that will require remarkable proof and execution.”
Dojo offers various advantages, according to the analysts, including in that it lets Tesla train its models with simulated data as well as real-life data. That’s critical in order to teach cars to deal with “one-off edge cases.”
Additionally, the analysts highlighted that Tesla’s Dojo efforts could help the company expand beyond the automotive sector in time.
“Dojo is designed to process visual data that can lay the foundation for vision-based AI models such as robotics, healthcare and security,” they wrote. “In our view, once Tesla makes headway on autonomy and software, third party Dojo services can offer investors the next leg of Tesla’s growth story.”