Silicom launches edge AI product line and announces partnership with Hailo

Silicom Ltd., a leading provider of networking and data infrastructure solutions, today announced the first in a new line of Edge AI products, the first fruit of a new partnership that it has formed with Hailo, a leading Artificial Intelligence (AI) chipmaker.

The new product line integrates Hailo’s AI-accelerators into Silicom’s existing Edge platforms, thereby solving performance challenges for Edge AI use cases. As a result, Silicom’s products will be able to offer visual processing and AI inference at the edge at a uniquely attractive price/performance ratio.

“We are excited to bring this powerful new Edge AI solution, our first AI-focused product, to the market,” commented Liron Eizenman, Silicom’s CEO. “By integrating Hailo’s groundbreaking chips into our products, we are making it cost-effective to move behavior analytics, human intrusion detection, facial recognition and vehicle analytics to the edge. This will help our customers unlock new use cases and provide new revenue streams and more value to their customers. In fact, we are already engaged in POCs and in discussions with customers on different use cases.”

Hailo’s Chief Technology Officer Avi Baum added, “By integrating Hailo’s top performing AI accelerators into Silicom’s Edge platforms, we can deliver unmatched AI inference performance, enabling real-time decision-making and advanced object detection and classification at the Edge. The wide range of Silicom Edge products empowered by the Hailo AI accelerators provides coverage for many Edge AI use cases.”

Mr. Eizenman continued, “Our partnership with Hailo is the first step in our AI journey and demonstrates our commitment to the Edge market, which we view as a primary growth driver. This partnership adds a whole new range of Edge use cases to the long and growing list of those we already address. We are excited about the opportunity that is being created by the meteoric growth of AI – and especially its need to solve training and inference networking problems in order to fully deliver on its promise.”

CT Bureau