Sterlite Technologies Ltd. (STL) rolled out its Neuralis suite of connectivity products designed for AI-driven data centers.

The Neuralis platform is positioned as a full connectivity foundation for modern data centers, which are increasingly being built around graphic processing unit (GPU)-heavy architectures and AI training environments. STL described the offering as a “central nervous system” for these facilities, designed to handle the higher density and bandwidth requirements tied to AI and hyperscale computing.

The announcement focuses on two main components.

The first is aimed at what STL calls “maximizing AI whitespace,” or the usable space inside a data center. Neuralis uses high-density, pre-terminated fiber cabling to support the large number of fiber connections required by GPU clusters. By completing fiber terminations in the factory rather than onsite, STL said it can reduce installation complexity, lower labor risk, and speed deployment timelines.

The second component targets data center interconnect (DCI), which connects facilities and moves large volumes of data between those locations. STL’s Neuralis portfolio includes its Celesta IBR cable series, which supports up to 6,912 fibers in a compact form factor. These cables are designed to handle the thermal and performance demands of AI environments while enabling high-capacity data transfer across data center campuses.

The company said the platform is designed to address a shift in how data moves across modern data centers. Instead of traditional “north-south” traffic – data flowing in and out – AI workloads are driving more “east-west” traffic, or data moving between servers within the data center. Neuralis is built to support that change with higher-density and higher-speed connectivity.

STL is also emphasizing its vertically integrated manufacturing approach. The company said it controls the full production process, from creating glass preforms and drawing fiber to cabling and final connectorization.

STL Managing Director Ankit Agarwal noted in the release that AI infrastructure requires “a level of precision and density that traditional cabling simply cannot meet,” adding that the company is shifting complexity away from onsite deployment and into controlled manufacturing environments.

The U.S. launch also highlights STL’s regional manufacturing footprint, including operations in South Carolina, which the company said will help support hyperscale and colocation providers building out AI infrastructure in North America.