#China #Hengqin #Montage Technology #CXL 3.2 MXC #Memory Controller

Montage Technology Unveils First Trial Production of CXL 3.2 MXC Chip

In a significant milestone for the tech industry, Montage Technology has announced the trial production of its CXL® 3.2 Memory eXpander Controller (MXC) chip, the first of its kind in the industry. Scheduled for production on July 31, 2026, this innovative chip is engineered to enhance memory expansion and resource pooling, particularly in the realms of AI and cloud computing.

Advantages of CXL 3.2 MXC Chip

The CXL 3.2 MXC chip aligns with the CXL Type 3 specifications, integrating both CXL.mem and CXL.io protocols. It comes equipped with PCIe® 6.x and CXL 3.2 protocol support, featuring impressive data transfer speeds of up to 64 GT/s. Additionally, the chip uses dual DDR5 memory controllers that can manage up to 8000 MT/s DDR5 memory. By efficiently transforming host-side CXL memory requests into DDR commands in real-time, this technology aims to transcend the memory limits of traditional server architectures.

This MXC chip not only incorporates a PCIe 6.x PHY and DDR5 PHY but also doubles as a platform for advanced RISC-V processor subsystems and a variety of management interfaces. Its versatility extends to several CXL memory expansion formats, including PCIe AIC cards and EDSFF modules, thus catering to emerging needs like memory pooling and tiered memory architectures.

Strategic Importance

Stephen Tai, President of Montage Technology, emphasized the importance of this launch, stating that it underscores the company's ongoing commitment to CXL innovation and commercialization. “With the increasing demand for large-capacity memory solutions in the era of AI, we are eager to help our customers achieve high-bandwidth and reliable memory infrastructures,” said Tai.

Notably, Montage Technology is not merely stopping at hardware; they are also offering a complete software development kit (SDK) along with comprehensive testing tools. These resources are designed to expedite product development, ensuring compatibility and speeding up mass production phases. Several leading memory manufacturers, including Samsung and SK hynix, have already validated the CXL 3.2 MXC, paving the way for eventual integration into their next-generation products.

Support from Industry Giants

Feedback from major industry players highlights the chip’s potential impact. Jangseok Choi from Samsung Electronics remarked, “The integration of CXL 3.2 MXC with DDR5 technology is a game-changer for data center customers, bridging the gap between performance and capacity.” Similarly, Uksong Kang from SK hynix noted that this advancement is critical for improving both memory expansion capabilities and resource sharing in next-gen applications, particularly in the context of generative AI.

Intel’s Dr. Debendra Das Sharma acknowledged the changing landscape of server memory architecture driven by AI and cloud technologies. He stated, “Montage Technology's CXL 3.2 MXC encapsulates the maturing CXL ecosystem, delivering the necessary protocol support to unlock next-generation Intel Xeon platforms.”

Amit Goel from AMD echoed this sentiment, noting that as data center workloads become increasingly demanding, the need for efficient memory architectures is more critical than ever.

Upcoming Events and Demonstrations

Montage Technology will officially showcase the CXL 3.2 MXC at two notable events: CXL DevCon and FMS (Future of Memory and Storage) in Santa Clara, California. Customers, industry peers, and partners are urged to engage at their exhibition booths to explore collaborative opportunities and gain deeper insights into this revolutionary product.

• - CXL Mini DevCon

Date: August 3, 2026
Location: Santa Clara Marriott, Santa Clara, USA

• - FMS

Date: August 4–6, 2026
Location: Santa Clara Convention Center, Booth #845, Santa Clara, USA

In summary, the CXL 3.2 MXC chip is set to transform how data centers operate and is poised to significantly enhance performance and productivity within the rapidly evolving AI landscape.