This brings the amount raised by Flex Logix since its formation in 2014 to $82 million, said Geoff Tate, CEO.
The money will be used to expand sales and marketing globally as well as to prepare for next-generation developments. Tate said that the embedded-FPGA said of the business is cash-flow positive and that the additional money would be used to propel the AI inference side of the business forward.
"We have a working chip. We need to build a sales team." Tate said the company's staff count stands at 45 but that he plans to take the company to 100 people by the end of 2021.
The funding round was led by Mithril Capital Management with participation by existing investors Lux Capital, Eclipse Ventures and the Tate Family Trust. Flex Logix's inference architecture is based on the rapid reconfigurability of FPGAs. It combines numerous one-dimensional tensor processors with reconfigurable, high bandwidth, non-blocking interconnect. The connections between compute and memory are reconfigured in millionths of a second as the model is processed. This architecture is the basis of Flex Logix’s InferX X1 edge inference accelerator which is now running YOLOv3 object detection and sampling to lead customers.
Peter Hebert, co-founder and managing partner of Lux Capital said he is confident Tate can build Flex Logix into a stand-alone public company. Tate said X1 chips and boards will be available for mass production in mid 2021 along with the availability of the InferX compiler, which takes Tensorflow Lite and ONNX neural network models and generates the code to run InferX X1.
Related links and articles: