According to CNBC, Tesla has purchased DeepScale, a Silicon Valley startup that employs low-wattage processors usually found in automotive applications to power precise computer vision. There are no details on how much Tesla spent on this acquisition.
Tesla has been short of autopilot engineers lately, and this move is made not only to strengthen the carmaker’s autonomous drive technology but also to employ DeepScale engineers’ abilities to plug the leak left by the workforce Tesla has lost.
An automaker with the resources, and a startup with the know-how: Tesla and DeepScale will create a new framework for autonomous vehicle driving together.
The American maker of electric cars is working hard to install “full self-driving hardware” which will handle all driving duties without human assistance. 3.0 Autopilot hardware is now being installed on all cars leaving the factories. Additionally, Tesla is calling upon EV owners to come to their facilities and get the upgrade.
Tesla had announced earlier that is would offer full self-driving capability in its cars by the end of this year. This statement seems unlikely to make it to reality without the help of DeepScale. With the 3.0 Autopilot hardware available, only the autopilot software is required for full autonomy. Tesla did release Version 10 of the software recently, named Smart Summon. However, it does not offer fully autonomous driving.
There have been news of Tesla cars failing to follow protocol while driving autonomously in Smart Summon mode, with owners reporting near-misses and cars failing to apply brakes at the right time in some cases. This is where DeepScale comes in, to rectify existing software glitches, and build a comprehensive response framework for the autonomous car to apply in real-world driving.
DeepScale has been dabbling into artificial intelligence that could solve problems or create new standards in the nascent autonomous driving industry.
Even when Tesla releases the desired software, the chances of it being fully autonomous are very low. The DeepScale acquisition is so fresh that the new team needs at least another year to come up with a substantially advanced piece of technology such as this. Time will tell how Tesla’s latest purchase directs the course of development within the firm.