AIStorm and Tower Semiconductor Introduce Cheetah HS Imager with On-Chip Neural Network

The Cheetah HS imager, developed by AIStorm and Tower Semiconductor, is a high-speed, 120-megapixel imager with first-layer AI capability, capturing up to 260,000 frames per second, 2,000 to 4,000 times faster than conventional CMOS sensors. This breakthrough technology enables slow-motion vision at a fraction of the cost and power consumption of competing high-speed cameras. The Cheetah HS is designed for applications such as robotics, drones, security and surveillance tracking systems, manufacturing and assembly lines, and biometric unlock systems.

Key Takeaways:

  • The Cheetah HS imager captures up to 260,000 frames per second, 2,000 to 4,000 times faster than conventional CMOS sensors.
  • The imager is designed to provide slow-motion vision at a fraction of the cost and power consumption of competing high-speed cameras.
  • The Cheetah HS features a charge-domain architecture that converts incoming photons to charge, computes the first neural-network layer in analog form, and outputs a pulse train that can be processed by downstream networks.
  • The imager has a programmable capture rate, allowing developers to optimize frame rates for specific applications.
  • The Cheetah HS imager is available now in both chip form and full reference-camera systems.

Statistics:

  • The Cheetah HS imager can capture up to 260,000 frames per second.
  • The imager is 2,000 to 4,000 times faster than conventional CMOS sensors.
  • The Cheetah HS features a built-in LED driver that can output up to 40 mA of current.
  • The imager has a low power consumption and a high-performance architecture that makes it suitable for a wide range of applications.

Sources:

  • GlobeNewswire report, August 12, 2025
  • AIStorm press release, August 12, 2025
  • Tower Semiconductor website, towersemi.com
  • Contify.com, The MIL Network (http://milnz.co.nz/mil-osi-aggregation/), 2025.