Startup Hailo announced on Thursday it has secured $60 million in funding for the development of its edge AI chip for inference, VentureBeat reported. The Hailo-8 has 26 TOPS of performance and as an edge chip is aimed at applications such as automotive, smart cities and robotics.
The $60 million was raised in a series B round. Hailo’s CEO says it will be used to accelerate the development of its AI chip for the edge, called the Hailo-8. Hailo-8 is supposed to ship early this year and has been sampling to select customers for over a year, indicating that it is close to launch.
It is also in the process of obtaining the ASIL-B safety rating at the chip level and the most stringent ASIL-D (used for instance in automotive) at the platform level. “The new funding will help us [deploy to] … areas such as mobility, smart cities, industrial automation, smart retail and beyond,” Hailo’s CEO said. It is also aiming at fully autonomous vehicles, smart cameras, smartphones, drones, AR/VR platforms, and even wearables in the future.
The Hailo-8 has 26 TOPS of performance and an efficiency of 2.6TOPS/W, according to VentureBeat. In a benchmark, it performed similar to the 30W Nvidia Xavier AGX while only consuming 1.7W.
As an edge AI chip, Hailo finds itself in a crowded field with both established companies and startups:
- Huawei HiSilicon Ascend 310: 16 TOPS (2.0 TOPS per watt)
- Intel Keem Bay: ~20 TOPS (~6 TOPS per watt)
- Nvidia Jetson Xavier NX: 21 TOPs (1.4 TOPS per watt)
- Google’s Edge TPU: 4 TOPs (2 TOPS per watt)
- AIStorm: 2.5 TOPs (10 TOPS per watt)
- Kneron KL520: 0.3 TOPs (1.5 TOPS per watt)
Hailo claims it uses a “Structure-Defined Dataflow” architecture that consumes less power than other chips and doesn’t need active cooling. Its software supports Google’s TensorFlow framework and the open ONNX format for deep learning models.
The $60 million funding brings the total amount raised to $88 million.
