Google disclosed their AI chips in 2018 conference at San Francisco.
Google is going beyond building artificial intelligence (AI) chips for its data centers. At the Google Cloud, Next 2018 conference which happened in San Francisco announced two new products whose main aim will be helping customers to develop and deploy intelligent connected devices.
First Product is ‘Edge TPU,’ which is basically a hardware chip whose purpose would be built the application-specific integrated circuit (ASIC) and it will help to design to run Artificial Intelligence at the edge. This technology will bring the brain to devices at extremely low cost and high power efficiency without compromising performance.
TPU consist of Tensor Processing Unit which is an application for a specific integrated circuit which is designed by Google for accelerating machine learning workloads, which will be available at the edge.
Second Product is Cloud IoT Edge which is basically a software which will extend Google Cloud’s data processing for AI capability to connected devices.
Couples of years ago Google already announced that they will be adding their Tensor Processing Unit to their cloud platform which will help in accelerating workload and speed.
People using Google Cloud Platform can connect to Cloud TPUs from custom VM types to balance processor speeds, memory, and high-performance storage resources for running compute-intensive workloads.
People are expected to utilize Cloud TPUs for training Machine Learning models. Google has introduced an Edge TPU development kit for developers, which will be available from October. This kit will include a system on module (SOM) that combines Google's Edge TPU, an NXP CPU, Wi-Fi, and Microchip's secure element in a compact form factor.