SANTA CLARA, Calif., Aug. 6, 2025 -- XCENA™, a South Korea–based startup, unveiled its first product, MX1 computational memory, this week at the Future of Memory and Storage (FMS) event. XCENA is displaying MX1 at FMS (booth #734) and will be offering working samples to select partners starting in October 2025. A production-ready version is planned for 2026.

CXL, The Foundation of MX1's Architecture

MX1 is the world's first computational memory product supporting PCIe® 6.0 and Compute Express Link® (CXL®) 3.2, the latest version of CXL, an open standard enabling high-speed, low-latency communication between CPUs, memory, and accelerators. MX1 brings compute capabilities closer to the memory layer to overcome the limitations of traditional memory hierarchies. This reduces the overhead of moving large volumes of data between memory and the CPU and allows data to be processed near where it resides — an architectural shift that improves overall system performance and reduces total cost of ownership (TCO) in data centers.

CXL introduces a more dynamic and flexible approach to system design. By supporting memory expansion, cache coherency, and direct memory access, the standard allows more efficient coordination across heterogeneous components. These capabilities form the foundation that MX1 builds on in order to scale data processing and reduce latency in demanding workloads.

Under the Hood: How MX1 Works

MX1 utilizes near-data processing, incorporating thousands of in-house RISC-V cores within the chip to handle computation near memory. This approach minimizes CPU involvement in data movement, accelerating tasks such as vector database operations, data analytics, and memory-intensive queries.

MX1 provides SSD-backed memory expansion, delivering petabyte-scale capacity with minimal latency overhead. It also incorporates advanced reliability features and hardware-assisted compression to maximize usable capacity.

Software Support and Developer Integration

To maximize the benefits of MX1, software support is essential. XCENA provides a fully-integrated software development kit (SDK) comprised of low-level drivers, runtime libraries, and supporting tools to help developers evaluate, prototype, and deploy MX1 with minimal integration effort.

The software stack ensures compatibility with standard system environments and development workflows, enabling seamless adoption across data-intensive applications, from in-memory analytics to AI inference acceleration.

As MX1 enters its proof-of-concept phase, XCENA invites developers, system architects, and partners to explore the new possibilities enabled by computational memory. Together, the industry is moving one step closer to a more flexible and efficient performance computing paradigm.

XCENA Wins Memory Innovation Award at FMS 2025

This week, XCENA was awarded the "Most Innovative Memory Technology" award at FMS for its breakthrough innovation that addresses the "memory wall" in modern computing systems. The award marks the company's second consecutive recognition at the event, following its selection as "Most Innovative Startup" in 2024.

"Computational Memory represents an emerging architectural approach that aims to accelerate performance and efficiency, particularly for data-intensive tasks. It does this by minimizing data movement between processing and memory components," said Jay Kramer, Chair of the Awards Program and President of Network Storage Advisors, Inc. "We are proud to recognize XCENA with MX1, the world's first computational memory controller supporting CXL 3.X and enabling high-throughput acceleration for data intensive workloads like AI and analytics."

FMS 2025 is currently taking place in Santa Clara, California, from August 5-7, bringing together global leaders in memory and storage technology. As an exhibitor, XCENA is showcasing MX1 in booth #734 and engaging with prospective partners, customers, and developers.

About XCENA
Founded in 2022, XCENA is a fabless semiconductor company developing intelligent memory solutions and data-centric computing architecture with a foundation built on Compute Express Link® (CXL®), an open coherent interconnect technology. The company focuses on improving performance and efficiency in data-intensive systems by pushing the boundaries of memory technology. To learn more about XCENA or its products, visit www.xcena.com and follow us on LinkedIn.

Compute Express Link and CXL are trademarks of the CXL Consortium.

Media Contact
contact@xcena.com