Rambus Inc. (NASDAQ: RMBS) announced the release of its PCIe 7.0 Switch IP with Time Division Multiplexing, designed for AI and data center infrastructure applications. The new intellectual property aims to address bandwidth, latency, and scalability requirements in AI, cloud, and high-performance computing systems.

The switch IP incorporates time division multiplexing capabilities to schedule and multiplex traffic across shared links. The technology is built on the PCIe 7.0 specification and targets next-generation AI and data center system-on-chips that require high bandwidth density and traffic management capabilities.

"The acceleration of AI is fundamentally reshaping system architectures, and it's no longer sufficient to simply add more lanes or more endpoints," said Simon Blake-Wilson, senior vice president and general manager of Silicon IP at Rambus. "With our PCIe 7.0 Switch IP with TDM, Rambus is giving system architects a new degree of freedom to scale bandwidth efficiently and deterministically."

The new switch IP joins Rambus' existing PCIe 7.0 IP portfolio, which includes controllers, retimers, and debug solutions. The company stated the technology is designed to integrate into ASIC platforms and support disaggregated and pooled compute architectures.

According to the press release, the switch IP enables more efficient utilization of PCIe links and supports data movement across CPUs, GPUs, accelerators, and NVMe storage in AI infrastructure systems. The technology is intended for use in AI training, inference, and data movement applications.