Lumos Infinitas

Lumos InfinitasLumos InfinitasLumos Infinitas
Home
Products
About
Contact
Career

Lumos Infinitas

Lumos InfinitasLumos InfinitasLumos Infinitas
Home
Products
About
Contact
Career
More
  • Home
  • Products
  • About
  • Contact
  • Career
  • Home
  • Products
  • About
  • Contact
  • Career

Optical Engine for Pluggable

 Our Optical Engine for Pluggable modules is a high-performance silicon photonics platform designed to enable next-generation AI clusters, cloud infrastructure, and hyperscale data-center networks. Engineered for high bandwidth, low power consumption, and high-volume manufacturability, the optical engine delivers reliable performance for next-generation optical transceivers.
Leveraging advanced silicon photonics integration, the platform combines high-speed optical modulation, photodetection, and passive photonic components into a compact optical engine optimized for pluggable transceiver applications. 

Key Features

  •  High-performance silicon photonics optical engine
  • Optimized for AI and hyperscale data-center interconnects
  • Low power consumption and high bandwidth density
  • Scalable architecture supporting multiple generations of Ethernet standards
  • Compact footprint for next-generation pluggable modules
  • Designed for high-volume manufacturing and reliability

Supported Data Rates

  • 100G per lane
  • 200G per lane

Supporting total module capacities

  • 400G
  • 800G
  • 1.6T

Available Configurations

  • DR4
  • DR8
  • 2×DR8
  • FR4
  • 2×FR4

Applications

  • AI accelerator clusters
  • Hyperscale data centers
  • Cloud networking
  • High-performance computing (HPC)
  • Ethernet optical interconnects

TRx Optical Engine for NPO

 The TRx Optical Engine for Near-Packaged Optics (NPO) is a high-density silicon photonics transceiver engine designed to enable next-generation AI networking and high-performance computing systems. By placing the optical engine in close proximity to the switch or compute ASIC, the NPO architecture minimizes high-speed electrical channel losses while significantly improving bandwidth density, signal integrity, and power efficiency.
Our integrated transmit/receive (TRx) optical engine combines high-speed optical transmitters and receivers into a compact, scalable package optimized for next-generation Ethernet and AI interconnects. 

Key Features

  • Integrated transmit and receive (TRx) optical engine
  • Silicon photonics platform optimized for NPO architectures
  • Ultra-high bandwidth density
  • Reduced electrical channel loss
  • Low power consumption
  • High-speed PAM4 optical transmission
  • Compact footprint for near-package integration
  • High reliability and manufacturability

Supported Data Rates

  • 100G per lane
  • 200G per lane

Supporting aggregate bandwidth

  • 400G
  • 800G
  • 1.6T

Available Configurations

  • DR4
  • DR8
  • 2×DR8
  • FR4
  • 2×FR4

Applications

  • AI accelerator clusters
  • Ethernet switch ASICs
  • AI scale-up and scale-out networks
  • Hyperscale cloud infrastructure
  • High-performance computing (HPC)
  • Optical fabrics for next-generation data centers

Optical Circuit Switch

 The SiPh Optical Circuit Switch is a high-performance silicon photonics switching platform designed to provide scalable, low-latency optical connectivity for AI infrastructure, hyperscale data centers, and high-performance computing (HPC). By establishing direct optical paths between compute nodes, GPUs, and network resources, the switch significantly reduces power consumption, latency, and network congestion compared with conventional electrical switching architectures.
Built on advanced silicon photonics technology, the optical circuit switch delivers high port density, low insertion loss, and fast switching performance while enabling flexible and energy-efficient optical fabrics for next-generation AI systems. 

Key Features

  • Silicon photonics optical circuit switching
  • Non-blocking optical switching architecture
  • Low insertion loss
  • Low-latency optical path establishment
  • High port density
  • Low power consumption
  • Scalable modular architecture
  • ​High reliability and manufacturability

Available Configurations

  • 32 × 32
  • 64 × 64
  • 128 × 128
  • 256 × 256
  • Custom scalable architectures available upon request

Applications

  • AI accelerator clusters
  • GPU-to-GPU optical interconnects
  • Optical fabrics for AI infrastructure
  • Hyperscale cloud data centers
  • High-performance computing (HPC)
  • Machine learning training clusters
  • Optical network disaggregation

Copyright © 2026 lumosus - All Rights Reserved.


Powered by

This website uses cookies.

We use cookies to analyze website traffic and optimize your website experience. By accepting our use of cookies, your data will be aggregated with all other user data.

Accept