New Tesla Cars Switch to Autopilot 2.1 And Have More Computing Power

Tesla Autopilot demonstration 1 photo
Photo: YouTube screenshot
Tesla Motors may have started as the company hell bent on convincing the world that electric propulsion is the way to go forward, but in recent years it added a new dimension to its ambitions.
Just like electric powertrains, Tesla isn't a lone industry figure as far as developing autonomous cars is concerned, but just like electric powertrains, it may very well be the most advanced of the bunch. Without decades of tradition to protect and with a visionary man at the helm, Tesla boldly went where no one had gone before and offered its Autopilot system way before other manufacturers were ready to take the risk.

Admittedly, the Autopilot is more or less capable of doing the same things as the driver aid systems in a Mercedes-Benz, but it's the way Tesla marketed the feature that made all the difference. It also led to the tragic demise of Joshua Brown, a man whose name will always be associated with the Autopilot, but the company wasn't legally at any risk, so the semi-autonomous system carried on unabated.

So much so that in October 2016, Tesla introduced the Autopilot 2.0. All vehicles produced from that date on were equipped with a new suite of hardware that, Elon Musk claimed, was enough to give the EVs Level 5 autonomy one day. But since Tesla had broken up with Mobileye and was forced to start over from scratch a few months before, the Autopilot 2.0 was actually less proficient than its previous iteration at first.

Frankly, it could be argued that it still is, which proves that replacing Mobileye's work turned out to be harder than anticipated. Whatever the case, sources from inside the company told Electrek that new Tesla cars are already being equipped with enhanced hardware meant to increase the vehicles' computing power.

Since the Autopilot relies heavily on video feed from its multitude of cameras (as opposed to other systems which use a LIDAR), the most stressed out components of the NVIDIA Drive PX2 supercomputer is its GPU. Which is why all Tesla models ordered today (S, X, and 3) allegedly come with one extra GPU, prompting the 'Autopilot 2.5' name that insiders use to refer to the new layout.

The same website obtained a comment from a Tesla Inc. representative who insisted it's not really such a big deal: “The internal name HW 2.5 is an overstatement, and instead it should be called something more like HW 2.1. This hardware set has some added computing and wiring redundancy, which very slightly improves reliability, but it does not have an additional Pascal GPU.”

Well, it may not be, but it comes at a time when Elon Musk postponed his planned coast-to-coast autonomous trip in a Tesla, so it might be that the company's engineers had underestimated the hardware needed to cope with Level 5 autonomy and are now taking the necessary steps to rectify it.
If you liked the article, please follow us:  Google News icon Google News Youtube Instagram
About the author: Vlad Mitrache
Vlad Mitrache profile photo

"Boy meets car, boy loves car, boy gets journalism degree and starts job writing and editing at a car magazine" - 5/5. (Vlad Mitrache if he was a movie)
Full profile


Would you like AUTOEVOLUTION to send you notifications?

You will only receive our top stories