Israel-based artificial intelligence (AI) chipmaker Hailo announced a $136 million Series C funding round led by Poalim Equity and Gil Agmon. Hailo Chairman Zohar Zisapel, ABB Technology Ventures, Latitude Ventures, OurCrowd, Carasso Motors, Comasco, Shlomo Group, Talcar Corp., and Machshiri Tnua also contributed to the funding round.
The Series C brings Hailo’s total raised to $224 million since it was founded in 2017. Former Intel SVP Mooley Eden joined Hailo’s board of directors and former Mellanox CEO Eyal Waldman joined its advisory board as part of the latest round.
According to Hailo CEO Orr Danon, the funding will address surging interest in the company’s first AI processor, the Hailo 8, and allow the company to develop a new generation of AI products.
“This, from my perspective, is our 8086,” he said, evoking Intel’s famous x86 processor, which spurred one of the industry's most prolific processor families.
Like Intel's 8086, Danon believes Hailo 8 is the beginning of something much larger.
AI Processing In a Gum Stick Form FactorThe Hailo 8 is designed for a variety of low-power, small-footprint edge applications including autonomous vehicles and industrial IoT. It’s available as a standalone chip for embedded applications or in an M.2 PCIe form factor, about the size of a stick of gum.
Despite its diminutive size and power requirements, the Hailo 8 is capable of 26 trillion operations per second, features integrated memory, and has a thermal design power of between 3 watts and 5 watts depending on the workload.
Additional chips can be added to a system, and performance will scale linearly with power and heat, Danon added.
The chip is particularly well suited to video analytics workloads, like object detection, but isn’t limited to that use case, he explained, adding the chip is completely programmable and can run a wide range of both off-the-shelf and custom AI inference models to suit customer demand.
Hailo 8 began sampling in the first quarter, and the company is now working with more than 100 customers to implement the small form factor AI accelerator.
Comments