Shares in Tesla went on a bit of a burn this week – adding more than $A125 billion in market value a single day on Monday as the stock jumped 10 per cent after an important upgrade from one of the car industry’s leading analysts.
The Morgan Stanley led by Adam Jonas has for many years provided some of the best insights on the Tesla-led transition to EVs, and what it means for the broader car industry. But until now it has been largely skeptical about the prospect of robo-taxis and autonomous driving.
That all changed this week with the release of a detailed report on Dojo, Tesla’s in-house supercomputer, in what amounts to an extraordinary reassessment of the company’s potential worth.
Jonas has upped his target price for Tesla from $US250 a share to $US400 a share, which would translate into a market capitalisation of some $US1.3 trillion ($A2 trillion). It might get even better than that, because Jonas also has a “bull case” stock price target of $US550 a share.
The company’s share price responded on Monday by jumping to a closing price of $US273 a share.
Why the change? It’s all about Dojo, the in-house supercomputer that Morgan Stanley says could turbo-charge adoption rates in robotics-taxis, something the investment bank had been cautious about till now.
It will also boost what’s known as network services, which is mostly about licensing its hardware to other companies and includes autonomy, charging, maintenance and upgrades.
But what is Dojo? And why the change?
Dojo is a purpose-built supercomputer designed in-house by Tesla to train its computer vision systems. It is often associated with the Optimus Humanoid but is also largely being driven by Elon Musk’s obsession with Tesla’s Full Self Driving (FSD) technology.
Tesla is spending at least $US1 billion a year on its development. More importantly, it has hundreds of million of miles of data from Tesla EVs driving around the world.
“Tesla cars are sensor encrusted robots making life and death decisions in highly unpredictable environments and situations,” the investment bank team writes.
“Tesla’s ability to improve the efficacy of its full self driving (FSD) system is limited by the ability to collect, label and process real world video data from the edge and to train these robots from the experience of its fleet in service, which is 5mm units today and closer to 50mm by end of decade.”
It says Tesla has put together a highly experienced semiconductor team, and has built a custom AI ASIC chip that can operate more efficiently (in terms of energy consumption and latency) than the leading cutting-edge general-purpose chips on the market from Nvidia, and potentially at a fraction of the cost.
“We believe Dojo can represent the next step-change in market perception of Tesla,” Morgan Stanley writes.
“Dojo emphasizes 3 of Tesla’s core capabilities: 1) speed, 2) performance, and 3) cost. In the near term, we believe Dojo can accelerate the development and monetization of Tesla’s software and services business.
“Longer term, we see scope for Dojo to provide avenues for Tesla’s software and hardware capabilities to extend well beyond the auto industry. If Dojo can help make cars ‘see’ and ‘react,’ what other markets could open up? Think of any device at the edge with a camera that makes real-time decisions based on its visual field.”
Indeed, it is what Jonas and his team describe as the “Muskonomy”, because it will have applications for Twitter/X, SpaceX and the battery business.
As for EVs, the development of Dojo puts Tesla at the core of the global industry, where rival car makers depend on it for licensing for charging and FSD hardware, and battery know-how.
“The more we looked at Dojo, the more we realised the potential for underappreciated value in the stock,” the analysts write.
“Tesla is not the first tech player to attempt to build a custom silicon system in-house. What’s unique about Tesla is the company’s longtime experience with advanced driver assist systems (ADAS).
“It has commercialized a vast network of vehicles that is constantly increasing (400k+ FSDs on the road already collecting data from 300+ million miles traveled). In addition, the company has brought together a world class design team, and has allocated expansive resources towards the autonomy problem.
“While it is difficult to explicitly validate the many claims Tesla has made about Dojo’s cost and performance, we believe Tesla has a chance of bringing forth a competitive customized solution given the company’s innovation track record and capabilities.”
Giles Parkinson is founder and editor of The Driven, and also edits and founded the Renew Economy and One Step Off The Grid web sites. He has been a journalist for nearly 40 years, is a former business and deputy editor of the Australian Financial Review, and owns a Tesla Model 3.