More Efficient AV Brains & More #CES2019

A hardware solution that allows Artificial Intelligence algorithms to run more efficiently is what Thinci promises with its Graph Stream Processing (GSP) computing architecture. Shawn Holiday, Thinci’s Senior Director of Product Marketing, explains that Thinci’s founders, who are veteran GPU designers, devised a new approach to graphics processing; one that reduces the need for memory and lowers clock speed, relative to traditional GPUs.

The upshot is that their offering promises to significantly reduce power consumption and increase efficiency; which is a big deal for products in power-constrained applications, such as data centers, edge devices (e.g. video cameras) or autonomous vehicles with electric drivetrains.

Holiday indicates that it is easy for OEMs to port their autonomous vehicle software stacks to Thinci’s graphics processing platform. What excites the auto industry, according to Holiday, is the promise of reducing the size and the power consumption of the brains of autonomous vehicles from 2,000 to 3,000 Watts to a few hundred Watts.

Their latest oversubscribed round of investment gave this fabless semiconductor company a major stamp of approval from major automotive OEMs, such as Daimler  DENSO, and, indirectly, Toyota through the Mirai Creation Fund.  Denso and Toyota’s announcement yesterday of a new joint venture for in-vehicle semiconductors, as well as Denso’s recent investment in Uber’s autonomous vehicle division,  have got to be more positive developments for Thinci and its GSP technology.

[Note: In November 2019, Thinci changed its name to Blaize].