The introduction of an expanded portfolio of modules enables developers to deliver a single, scalable architecture that can power everything from low-power devices to enterprise edge infrastructure.
As AI moves from cloud environments to factories, transportation systems, robotics, and intelligent infrastructure, developers face increasing challenges related to power consumption, memory bandwidth, deployment complexity, and infrastructure costs. MemryX Inc., a fabless semiconductor company that develops edge AI processing solutions that leverage proprietary compute-at-memory technology, today expanded its family of cascade deployment platforms with three new products built on the MX3 accelerator architecture, enabling AI deployment across embedded systems, edge devices, and dense edge server infrastructure. These products will be unveiled this week at Automate 2026 in Chicago’s McCormick Place North Building, booth 15045.
MemryX Cascade 100P PCIe Accelerator
MemryX Cascade 100P PCIe accelerator is designed for high-density edge server deployments and demanding multi-camera AI applications. The card is built on a 16-chip array of MX3 processors, enabling parallel inference of hundreds of real-time video streams. With its single-slot passive cooling form factor, the Cascade 100P can be easily integrated into standard server environments, allowing operators to maximize compute density without adding complex thermal or power budget constraints. This makes it an ideal solution for smart manufacturing, intelligent transportation, video analytics, and multi-camera AI deployments.
MemryX Cascade 100U USB Accelerator
MemryX Cascade 100U USB Accelerator is designed to bring AI acceleration to existing systems through standard USB Type-C connectivity. Featuring two MX3 accelerators, the device allows developers to quickly evaluate, prototype, and deploy AI vision applications without redesigning existing hardware platforms.
Also read: AiThority interview with Matej Bukovinski, Chief Technology Officer at Nutrient
MemryX Cascade 100R Raspberry Pi HAT+
MemryX Cascade 100R Raspberry Pi HAT+ is designed to bring hardware-accelerated AI inference to the Raspberry Pi ecosystem for robotics, machine vision, education, and low-power industrial applications. Built around two MX3 processors connected via PCIe, the HAT can be cascaded to support large-scale AI deployments.
All three products are supported by the MemryX Developer Hub, which provides a common software environment for compiling and deploying AI models across the MX3 product family.
“AI is no longer a model problem, it’s a deployment problem,” said Ross Jatou, CEO of MemryX. “Developers and system builders are using Raspberry Whether you’re prototyping on the Pi, adding AI acceleration to an existing system over USB, or deploying hundreds of video streams to edge servers, you need a common architecture that scales seamlessly and delivers performance without the power and memory penalties of traditional approaches. Our expanded Cascade platform family provides that foundation, allowing customers to go from development to production on a single, scalable, highly efficient AI architecture powered by MX3.”
Together, these new products establish a complete MX3 deployment ecosystem, allowing organizations to scale AI applications from development and embedded deployment to dense edge server environments using a common hardware and software foundation.
