Ambarella, Inc. is an American semiconductor company headquartered in Santa Clara, California. It develops systems on a chip capable of processing images, compressing video and running artificial intelligence algorithms within devices. It operates as a fabless chip designer: the company designs its components’ architecture and software environment, while manufacturing is outsourced to industrial partners. Its corporate website is ambarella.com.
From video compression to computer vision
Founded in 2004 by Fermi Wang, Les Kohn and others, Ambarella initially focused on high-definition, low-power video processing. Its components found markets in action cameras, surveillance cameras and other video capture devices. The company has been listed on Nasdaq since 2012 under the ticker symbol AMBA.
Its evolution reflects the shift from simply capturing images to interpreting them automatically. In 2015, Ambarella acquired VisLab, a company that grew out of research conducted in Parma, Italy, specializing in computer vision and autonomous driving. This acquisition strengthened its expertise in environmental perception, a field that has become central to its automotive development work.
Components for on-device image analysis
Ambarella’s portfolio combines image signal processing, video encoding and neural network acceleration. Its CVflow architecture aims to perform computer vision functions with power consumption suited to the constraints of embedded equipment. The goal is to enable a camera or vehicle to detect and analyze elements in its surroundings without systematically transmitting all data to a remote server.
Target markets include video surveillance, security devices, advanced driver assistance systems and certain robotics applications. In the automotive sector, requirements include making use of camera images and developing computing architectures that combine multiple sensors. Ambarella supplies components and software tools to manufacturers and equipment suppliers; it does not itself sell vehicles or complete surveillance systems.
What next?
Ambarella’s trajectory depends on the adoption of artificial intelligence at the network edge, as close as possible to the sensors. Local processing can reduce latency, limit data transfers and help meet certain privacy requirements. However, these advantages must translate into measurable improvements in cost, power consumption and performance for customers.
The company operates in a competitive environment that includes other computer vision specialists and major computing platform providers. In the automotive sector, qualification and production launch cycles are long. The challenge will therefore be to turn its technological developments into sustainable industrial programs while keeping pace with rapidly evolving artificial intelligence models and safety requirements.