Telescent adds high-density optical circuit switch for DR4/DR8 GPU links

Telescent has introduced a high-density robotic cross-connect system aimed at DR4 and DR8 parallel-optics interconnects in AI GPU training clusters, positioning it as an Optical Circuit Switch (OCS) and automated fiber patch-panel option for large-scale GPU fabrics. The company says the system is designed to help operators automate turn-up and reconfiguration of fiber paths as cluster fiber counts grow, with the goal of reducing manual patching errors, downtime, and operational overhead.

According to Telescent, the new offering extends its G5 robotic platform to support the high fiber counts typical of AI cluster topologies. The company frames the product around operational needs in GPU clusters that use DR4 transceivers and other parallel optics approaches, where the number of fiber connections can scale dramatically. Telescent claims the robotic cross-connect can manage “hundreds of thousands” of fiber connections in large-scale GPU fabrics, enabling automated reconfiguration and physical-layer changes without manual patching.

For data center engineers, the core idea is physical-layer automation for parallel optics links: instead of dispatching technicians to repatch high-count fiber fields during cluster builds, expansions, or topology changes, Telescent is pitching a robotic, software-controlled approach. In dense GPU environments, manual fiber work is slow and error-prone, and the company is explicitly tying this system to avoiding patching mistakes that can trigger outages during production AI runs.

“The bandwidth requirements of AI infrastructure are rewriting the rules of data center fiber management. A single AI cluster can require hundreds of thousands of fiber connections, and the move to parallel optics architectures like DR4 multiplies that count significantly,” said Anthony Kewitsch, CEO and Founder of Telescent. “Our new high-density robotic cross-connect system gives operators a powerful automated solution to manage this complexity to ensure maximum GPU utilization and operational efficiency while future-proofing the physical layer for the next wave of AI innovation.”

Telescent said it will show a live demonstration of the new system at the Optical Fiber Communication Conference (OFC) 2026 in Los Angeles, California, March 17–19 at Booth #607. More information is available at telescent.com.

Source: Telescent

Get Data Center Engineering News In Your Inbox:

Popular Posts:

Screenshot
Five AI data centers to reach 1 GW power capacity in 2026, new analysis shows
Near-Packaged-Optics--Rethinking-the-AI-Data-Center-Interconnect
Near-Packaged Optics: Rethinking the AI Data Center Interconnect
30cf-data-center-pr
Carrier launches AquaEdge 30CF chiller to boost data center cooling reliability and uptime
shine 的複本 的複本 - 36
GenerMotor launches stackable HVDC generator modules for AI data center power
Low-chill_graphic
HRL Low-Chill single-phase liquid cooling targets high-density GPU racks with low pressure drop

Share Your Data Center Engineering News

Do you have a new product announcement, webinar, whitepaper, or article topic? 

Get Data Center Engineering News In Your Inbox: