GUIDE

Very Large Scale Photonics (VLSP) Light Engine

Guide® redefines how we build the light sources that power co-packaged optics (CPO). Hundreds of lasers and photonics components on a single chip. We’ve taken the laser out of the assembly line and put it into the foundry—turning an expensive manually-integrated component into a scalable, software-defined light engine.

Capacity

Up to 100T

Tbps total switch bandwidth supported in one RU

Control

Software Defined

Real-time telemetry

Spectrum

16 Wavelengths

100/200/400+ GHz spacing, up to 64 wavelength roadmap

Technology

VLSP

Very Large Scale Photonics

Solving the Industry Challenge

Completing the Photonic Stack

For years, the industry has focused on the “silicon” in silicon photonics. Unleashing the photonic revolution requires a scalable, reliable, and integrated light source.

Until now, scaling optical I/O meant building “laser farms”—sprawling arrays of discrete external laser small form factor pluggable (ELSFP) modules that consume faceplate space, complicate thermal design, and introduce hundreds of points of failure.

Guide changes the math. By integrating high-power laser arrays into a compact module form-factor, Lightmatter creates a universal light engine that scales with the fabric, not against it.

Strategic Alignment

While the industry iterates on loose CWDM grids, Guide delivers a tight DWDM system capable of driving the most advanced optical networks in the AI data centers.

01 — Scalability

HIghest Optical Output Power

100 mw+ per fiber.

02 — Control

Software-Defined

CMIS-compliant interface with real-time telemetry and active stabilization. Guide manages its own physics to maintain signal integrity.

03 — Tuning

Hyper-Local

Energy-efficient heating within the multi-wavelength cavity. Independent tuning of colors without thermal coupling or spectral drift.

04 — Reliability

Semiconductor-Grade

Replaces epoxy-bonded components like lenses with integrated photonics.

05 — Compatibility

Universal Ecosystem

The gold standard for CPO, NPO, and OBO. Decoupling the light source from hot XPUs and switch chips to enable the next generation of pluggable and co-packaged solutions.

Technology

VLSP Advantages

Density, BOM and Assembly

Guide enables massive density by integrating hundreds of light sources into a single footprint. By moving away from discrete ELSFPs—where packaging and manual alignment drive the bulk of the cost—the Guide light engine dramatically simplifies the bill of materials. One chip enables seamless scaling from 4 to 64 wavelengths with zero increase in assembly complexity.

Hyper-Local Tuning

The Guide light engine allows for hyper-local, energy efficient heating directly within the laser cavity. This locality ensures that lasers are not thermally coupled, allowing for precise, independent tuning of 16+ colors simultaneously on a single PIC without affecting adjacent channels.

Self-Healing & Redundancy

VLSP laser tunability enables N+M redundancy schemes. If a primary laser fails, the system can autonomously tune a backup laser to the exact required frequency and “swap it in” via the integrated photonic system, ensuring exascale reliability.

High-Volume Manufacturing (HVM)

On-PIC (photonic integrated circuit) wavelength tunability allows every laser on the wafer to be identical at fabrication. Targets are “software-controlled” post-packaging, enabling a single-SKU HVM strategy where one module design supports multiple frequency grids, dramatically improving yield.

The Scaling Wall

144 ELSFPs is not a roadmap

As aggregate switch bandwidth doubles every generation, the reliance on pluggable, discrete laser modules creates an untenable faceplate crisis. Guide provides the only viable path to 800T networks.

Mechanical I/O Collision (Legacy)

Mechanical I/O Collision diagram showing faceplate limitations

Port Escape
Gridlock

In liquid-cooled AI racks, the front panel (faceplate) is the only escape for networking cables. 144 ELSFPs physically consume the entire faceplate area, leaving zero room for the high-density data ports required to move AI traffic.

The Data Faceplate (Guide)

Data Faceplate diagram showing Guide solution

800T
Ready

Integrating the light engine inside the chassis with an internal coldplate reclaims the entire front panel for networking. This enables a 144-port high-density grid in a standard one RU footprint—the only buildable path for 800T AI interconnects.