LDA Technologies Launches LDA Maze: Ultra-Low Latency 25G Layer 1 Fabric for High-Frequency Trading

Quick Summary

LDA Technologies has launched LDA Maze, a non-blocking Layer 1 fabric delivering ultra-low latency of 3.6ns to 4.1ns. Supporting up to 160 x 25G links, it enables capital markets and AI firms to scale high-density connectivity while significantly reducing physical rack space and power consumption.

How Does LDA Maze Solve High-Density Connectivity Challenges?

Modern trading environments face a critical bottleneck: the need for high-density 25G connectivity without increasing physical footprint. LDA Maze addresses this by offering 160 links in a single fabric, a density level previously unseen at these speeds. By utilizing ARC6 connectors, the platform provides a flexible interconnect between FPGA boards, NICs, and front-panel links.

• 160 links of 10/25G or 40 links of 100G.
• 3.6ns latency measured port-to-port.
• Reduced rack space and power requirements.

This architecture allows firms to future-proof systems by running 10G today and scaling to 25G without a complete hardware redesign, ensuring long-term infrastructure efficiency.

What Performance Metrics Does LDA Maze Deliver for AI and Trading?

Performance in capital markets is measured in nanoseconds. LDA Maze delivers a predictable 3.6ns to 4.1ns port-to-port latency, ensuring that high-frequency trading (HFT) firms maintain a competitive edge. Beyond finance, the fabric supports AI workflows and High-Performance Computing (HPC) by aggregating links to 40 x 100G.

"There's been a gap in the market for high-density, 25G-capable Layer 1 technology, and LDA Maze addresses it. It provides a flexible, ultra-low latency interconnect between components within the chassis: front-panel links, FPGA boards, NICs and other equipment alike, over ARC6 connectors - so one fabric can power devices for finance, HPC and AI workflows," said Vahan Sardaryan, CEO and Co-founder of LDA.

How Does This Impact Infrastructure Scalability?

The modular I/O cards (SFP, QSFP, and QSFP-DD) allow the link layout to match the specific build rather than being constrained by the chassis. This configurable workflow means firms can chain multiple LDA Maze units together to create even larger fabrics, supporting massive scale without sacrificing performance.

FF NEWS TAKE:

LDA Technologies is hitting a sweet spot in the market. As ultra-low latency becomes a requirement not just for HFT but for real-time AI processing, the ability to pack 160 links into a single fabric with sub-4ns latency is a significant technical achievement. This announcement moves the needle by proving that high-density 25G connectivity doesn't have to come at the cost of increased latency or massive power draws.