XKL Adds Low Latency Optical Links for Distributed AI

Sep 17, 2026
XKL introduced MediaLight data center interconnect capabilities for distributed AI and high performance computing. The systems provide Layer 1 optical transport with latency below 5 microseconds.

XKL introduced specialized data center interconnect capabilities for its MediaLight product family, the company said in a press release. MediaLight Xponder and Muxponder systems provide deterministic optical transport with latency below 5 microseconds for distributed AI and high performance computing clusters.

The 1RU systems use direct, cut through Layer 1 DWDM connections to avoid delays from multi hop routing. They support software configurable 100G QSFP28 and 400G QSFP-DD transport over dark fiber or third party open line systems, with uncompressed and bit transparent connectivity.

XKL says the design reduces packet jitter, queueing delays and GPU idle time during large dataset transfers. The MediaLight systems are shipping for AI interconnect and enterprise deployments, with management through a standard command line interface and open MIBs.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Enterprise AI Brief, Silicon Brief or Daily AI Brief.

Also, consider following us on social media:

Enterprise AI AI Chips & Datacenters AI Brief AI Brief (X)
Free newsletter

Enterprise AI Brief

Weekly report on AI business applications, enterprise software releases, automation tools, and industry implementations.

AI Token Webinar

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more

You may also like

Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.