AI startup Hailo today announced that it’s teaming up with Foxconn and system-on-chip provider Socionext to launch BOXiedge, an edge computing processing solution for video analytics. If the companies’ claims bear out, BOXiedge could deliver “market-leading” energy efficiency for AI inference, benefiting applications like industrial internet of things, smart cities, and smart medical.
BOXiedge is the successor to a mini server Foxconn teamed up with Network Optix to launch in January, which confusingly shares the same name. Unlike the previous server, this new BOXiedge can perform image classification, detection, pose estimation, and other tasks on footage from up to 20 cameras simultaneously thanks to SocioNext’s SynQuacer SC2AA chip and Hailo’s Hailo-8 processor, which features an architecture that consumes less power than rival chips while incorporating memory, software control, and a heat-dissipating design.
Under the hood of the Hailo-8, resources including memory, control, and compute blocks are distributed throughout the whole of the chip, and Hailo’s software — which supports Google’s TensorFlow machine learning framework and ONNX (an open format built to represent machine learning models) — analyzes the requirements of each AI algorithm and allocates the appropriate modules.
Hailo-8 is capable of 26 tera-operations per second (TOPs), which works out to 2.8 TOPs per watt. In a recent benchmark test conducted by Hailo, the Hailo-8 outperformed hardware like Nvidia’s Xavier AGX on several AI semantic segmentation and object detection benchmarks, including ResNet-50. At an image resolution of 224 x 224 pixels per inch, it processed 672 frames per second compared with the Xavier AGX’s 656 frames and sucked down only 1.67 watts (equating to 2.8 TOPs per watt) versus the Nvidia chip’s 32 watts (0.14 TOPs per watt).
The edge AI hardware market is anticipated to be worth $1.15 billion by 2023, and Hailo — which raised $60 million in March — is hoping to beat rivals to the punch. Startups AIStorm, Esperanto Technologies, Quadric, Graphcore, Xnor, and Flex Logix are developing chips customized for AI workloads. Mobileye, the Tel Aviv company Intel acquired for $15.3 billion in March 2017, offers a computer vision processing solution for autonomous vehicles in its EyeQ product line. Baidu in July unveiled Kunlun, a chip for edge computing on devices and in the cloud via datacenters. And Chinese retail giant Alibaba launched an AI inference chip for autonomous driving, smart cities, and logistics verticals in the second half of 2019.
Foxconn is one of Hailo’s first publicly disclosed customers after NEC and ABB Technology. Previously, the startup said it’s working to build Hailo-8 into products from OEMs and tier-1 automotive companies in fields such as advanced driver-assistance systems (ADAS) and industries like robotics, smart cities, and smart homes.