A lot is riding on Tesla's autonomous-driving software, and the company appears to be making progress.
Tesla believes it will be able to launch a self-driving robotaxi service in late 2025. Estimates of how much that service could be worth range from hundreds of billions of dollars to several trillion.
While Tesla hasn't completed a truly self-driving cab ride, the company is constantly improving its driver-assistance products with the help of artificial-intelligence computing. The latest iteration was widely released in December.
Self-driving technology is critical to Tesla stock, so Barron's took the latest version of the electric-vehicle maker's Full-Self Driving product out for a spin to help investors not exposed to the technology understand the state of play. The new version is impressive.
Barron's has tested Tesla's FSD product several times before. The first review was in September 2023, when I wrote that the car drove "like a mash of a teenager with a learner's permit and an octogenarian." It was slow and steady, but unsure about all the situations encountered in daily driving.
FSD worked, but it wasn't nearly as good as an experienced human.
The car kept improving with each updated version of the software, though it still requires human oversight 100% of the time. New versions tested in 2024 were able to handle turns and intersections that gave the car trouble in the past.
Now there is FSD version 13.2, which is available to owners whose vehicles have the required hardware. Its enhancements include "reduced photon-to-control latency," whatever that means. Supervised self-driving starting from a parking spot is another innovation -- a forward step that is easier to understand.
Version 13.2 is really good. The number of times I have to take over on the road is down substantially, and the driving is significantly better. FSD is starting to feel like a solid human driver. It navigates parked vehicles, construction, difficult left turns, speed bumps, and pedestrians with a surprising level of expertise, though I am still paying attention 100% of the time.
Others have noticed something similar. "Lucky version 13," wrote Truist analyst William Stein in a recent report. "Our first, second, and third reviews of this technology each revealed material weaknesses. Our recent test drive of FSD v13 was more impressive, requiring no interventions. Still, imperfections remain obvious and prevent us from recommending its use."
He rates Tesla shares Hold and has a $360 price target for the stock, which closed Friday at $410.44.
Stein sounds more skeptical than Barron's, though FSD certainly isn't perfect. I can't project when it will be much better than a human driver.
Elon Musk believes that will happen in early 2025. "That is just, unvarnished, our internal estimate," the Tesla CEO said on his company's third-quarter earnings conference call in October. "That's not sandbagging or anything else. Our internal estimate is Q2 of next year to be safer than humans."
FSD version 13.2 gives investors some hope that Musk, who is famously optimistic about self-driving timelines, will be right this time.
He needs to be. Through Friday trading, Tesla stock was up roughly 70% since the company's Oct. 10 robotaxi event. Those gains added more than $500 billion in market value.
Expectations that the robotaxi service is close, and will be profitable, are running high. How FSD advances throughout 2025 will go a long way toward determining how Tesla stock will do in the new year.