Tech news in 3 minutes

CEO Interview with James Regan of Oriole

29 d ago

Mark Goranson, CEO of EMASS (Nanoveu’s semiconductor division) with over 45 years in the industry, leads the company in commercializing ultra-low-power edge-AI system-on-chip (SoC) solutions. EMASS builds next-generation edge AI semiconductors enabling intelligent devices to sense, understand, and respond to their environments continuously without cloud reliance or sacrificing battery life. Their ECS-DoT platform brings real-time AI onto wearables, drones, industrial systems, and smart infrastructure. The core problem is that battery-powered edge devices must process sensor data in real time within tight power budgets. Conventional processors often waste energy on irrelevant inputs, rely on cloud connectivity or external memory, increasing latency and power consumption. ECS-DoT uses event-driven architecture, keeping devices in an ultra-low-power state until meaningful activity is detected, processing AI models entirely on-chip to eliminate cloud dependency and extend battery life. Key application areas include wearables (smart glasses, hearables, fitness trackers) enabling private voice interfaces and gesture detection; battery-powered drones and robotics for real-time onboard AI within sub-milliwatt power, improving endurance for inspection, agriculture, and surveying; and industrial IoT for local anomaly detection and event-driven wireless communication, reducing bandwidth and prolonging battery life. Customers are most concerned about battery life—whether for medical wearables requiring long-lasting or smaller batteries, or drones needing higher utilization rates. EMASS also enables AI for device operation itself, like improving drone flight dynamics, helping customers compete better. In a crowded edge AI market, EMASS differentiates by focusing specifically on continuous always-on intelligence within a sub-milliwatt envelope, unlike solutions designed for larger power budgets. They combine event-driven sensing, integrated memory, dedicated AI acceleration, and ultra-low power. New developments include scaling on advanced process nodes for multi-modal workloads (vision, audio, motion), improving developer experience with model zoos and IDEs, and deepening sensor and distributor partnerships to ease adoption. Customers typically engage via developer kits for evaluation, benchmarking power, and integrating with the SDK, moving to proof-of-concept designs supported by field engineers and global distributors. Goranson emphasizes making the chip part of a working system for seamless adoption.