Tesla is acquiring tech startup DeepScale, a company that develops high-precision computer vision technology designed to run on low-wattage computer processors such as those found in most automotive applications, CNBC reported Monday.
Several of Tesla's Autopilot engineers resigned over the summer, and CNBC speculates that the DeepScale acquisition is as much a move for the company's technology as it is for its talent, which can help bolster the depleted team.
The acquisition deal included a position for DeepScale founder and CEO Forrest Iandola as a senior machine learning specialist working the Autopilot team, which is where DeepScale's technology—and the expertise of its staff—is expected to be put to use.
The other terms of the acquisition, including the purchase price, weren't disclosed.
The EV automaker is in the middle of a push to deploy its "full self-driving hardware" to new and existing models. Tesla models leaving the factory are now equipped with the 3.0 variant of the company's Autopilot hardware, which the company claims can deliver "full self-driving capability," which Tesla promised would arrive before next year, despite a lack of infrastructure and applicable laws.
Existing owners are also starting to receive hardware upgrades to make their older cars capable of full self-driving, however the necessary software has not yet been announced. Tesla just released Version 10 of its software suite earlier this week, without "full self-driving" capabilities. Smart Summon, which is a limited form of autonomous driving, is the closest customers can get to the real thing in the current software suite.
Unlike driver-less Smart Summon, Tesla's version of "full self-driving" will still require a human to be behind the wheel when it debuts. It's likely that it won't be "full self-driving" as advertised, but rather a very specific kind of driver assistance in limited cases.