CXL DevCon 2025: XConn Technologies Demonstrates Dynamic Memory Allocation Using CXL Switch and AMD Technologies

XConn Technologies announced a groundbreaking demonstration of dynamic memory allocation using Compute Express Link (CXL) switch technology at CXL DevCon 2025, taking place April 29-30 at Santa Clara Marriott hotel, CA.

The demonstration highlights a major advancement in memory flexibility, showcasing how CXL switching can enable on-demand memory pooling and expansion across heterogeneous systems.

The milestone, achieved in collaboration with AMD (Advanced Micro Devices Inc.), unlocks a new level of efficiency for cloud, AI, and HPC workloads. By dynamically allocating memory via the XConn Apollo CXL switch, data centers can eliminate over-provisioning, enhance performance, and significantly reduce TCO.

“Memory agility is the next frontier in computing, and this demonstration is a pivotal step toward delivering scalable, software-defined infrastructure for the most demanding AI and HPC environments,” said JP Jiang, SVP, product marketing and management, XConn. “With CXL-enabled dynamic memory allocation, we’re showing how memory can be pooled and distributed in real time, unlocking efficiencies that static architectures simply can’t deliver.“

At the core of the demo is Apollo CXL switch, an industry’s first to support both CXL 2.0 and PCIe 5.0 on a single chip. The switch enables TB-scale memory expansion with near-native latency and coherent memory access across CPUs, GPUs, and accelerators, including the 5th Gen AMD EPYC processors.

“CXL switching offers tremendous potential for the next generation of datacenter computing, especially in use cases like distributed shared memory, which can greatly enhance efficiency and reduce costs for data-intensive applications,” said Raghu Nambiar, corporate VP, data center ecosystems and solutions, AMD. “By combining AMD EPYC processors with the XConn Apollo CXL switch, XConn are helping to deliver on highly-flexible and adaptive memory infrastructure for next-gen data centers.“

Key benefits of XConn’s CXL-based dynamic memory allocation on display during event include:

• On-Demand Memory Pooling: Share and scale memory across systems to avoid over-provisioning.
• Low Latency Performance: Coherent access ensures memory behaves like local DRAM.
• TB-Scale Expansion: Ideal for AI inference KV caching, in-memory databases, and virtualization workloads.

Production samples of XConn Apollo XC50256 are available.

CXL DevCon 2025 attendees can experience the live demo at the XConn booth and learn more about the breakthrough in the technical session, Showcasing a CXL 2.0 Memory Pooling/Sharing System, presented by Jiang, Xconn, on April 29 at 2:40 p.m.