Tesla to acquire computer vision company DeepScale

Tesla has reportedly acquired DeepScale, an American start-up working on computer vision technology for self-driving cars. The company’s CEO joined Tesla as a machine learning expert in the Autopilot division.

Tesla and DeepScale have not disclosed anything about the acquisition, but CNBC says it has heard from various sources that the automaker has acquired the start-up. Forrest Land’s LinkedIn page shows that he now works at Tesla. Land is the CEO of DeepScale.

DeepScale was founded in 2015 and is working on computer vision systems for autonomous cars. The company’s software uses deep neural networks to interpret the environment. The goal of DeepScale is to run such software on relatively simple processors in cars.

The company is known for SqueezeNet, a neural network for computer vision that takes only 0.5MB and therefore needs to run on many systems. SqueezeNet was developed by DeepSpace and university researchers. SqueezeNet is available on GitHub and has a paper published on arXiv.

CNBC suggests that Tesla wants to attract talent with the acquisition, because many people have left the Autopilot team in recent times. According to The Information, there was discontent in the team because Musk was not satisfied with the progress being made. At least eleven employees, including veterans, are said to have left the company.