AIStorm

AIStorm is a semiconductor company developing AI-in-Sensor processors that operate at the edge. Founded in 2018 by David Schie, a seasoned executive with extensive experience in the semiconductor industry at companies like Maxim Integrated, Micrel, and Semtech, AIStorm is headquartered in Houston, Texas. Schie's background in analog design and founding of Linear Dimensions Semiconductor laid the groundwork for AIStorm's core technology.

The company's central business is the design and manufacturing of System-on-a-Chip (SoC) products that integrate artificial intelligence directly with sensors. This AI-in-Sensor approach processes raw, analog data directly from the sensor without first converting it to a digital format. By eliminating the digitization step, AIStorm's technology significantly reduces latency, power consumption, and system costs compared to traditional digital GPU-based solutions. This method allows for real-time processing directly at the edge, which is critical for applications requiring immediate action and analysis.

AIStorm's revenue is generated through the sale of its semiconductor chips to a diverse client base that includes manufacturers of mobile handsets, IoT devices, wearables, and advanced driver-assistance systems (ADAS). The company has cultivated strategic partnerships with its investors, who are also its customers, including leaders in biometrics, image sensors, and consumer electronics like Egis Technology, TowerJazz, and AsusTek. This model provides AIStorm with direct access to define solutions and technologies tailored to market needs. To date, AIStorm has raised a total of $29.2 million over two funding rounds: a $13.2 million Series A in February 2019 and a $16 million Series B in December 2020.

The product line includes several families of SoCs. The Mantis line is designed for continuous, always-on monitoring in applications like smart cameras. The Cheetah line targets high-speed applications such as facial recognition, while the Chameleon family provides an AI-friendly analog front-end for biometric and vibration sensing. These chips employ a charge-domain processing technique, which works with the charge directly from a sensor's pixels or a MEMS microphone, allowing for a monolithic design where the sensor and processor are on the same chip. This enables applications ranging from facial recognition and people counting to inventory monitoring and voice input for battery-operated devices.

Keywords: AI-in-Sensor, edge AI, analog computing, System-on-a-Chip, SoC, semiconductor, charge domain processing, edge computing, IoT processors, low-latency processing, computer vision, ADAS, biometric sensors, facial recognition, gesture recognition, voice input, occupancy sensing, machine learning at the edge, analog AI, low-power AI, smart sensors, image sensor processors, David Schie