Ultrafast photonics for next-generation compute
LightScale Photonics is building ultrafast photonic switching technology for AI interconnects. Our platform is designed for faster optical switching, lower communication power, and more adaptive network fabrics for next-generation compute.
AI infrastructure is hitting a communications wall
The bottleneck
As AI clusters scale, communication is becoming a first-order systems constraint. A 100,000-GPU training cluster needs on the order of 600,000 optical switches. That's roughly $600M in transceivers and 18 MW of power just for the network. Today's interconnect stack pays heavily in power, cost, latency, and fixed topology.
What's needed
Future systems need a switching layer that moves more data with less energy and reconfigures on workload time scales. Optical circuit switching has already cut interconnect power by ~40% vs co-packaged optics, 2.6× vs pluggables, and over 10× vs copper. The next step is to push significantly further.
Built for high-speed, low-energy switching
Our platform targets switching speed, energy, and footprint in regimes that conventional approaches struggle to reach.
Ultrafast switching
Push optical switching into a faster operating regime than conventional platforms allow.
Compact photonics
Enable denser integration and smaller photonic structures.
Lower switching energy
Target per-switch energy roughly 1,000× lower than current silicon photonic switches, by replacing heat-driven control with non-thermal switching paths.
Designed for AI training and inference
Target workloads include all-reduce, all-to-all traffic, MoE routing, and bandwidth-heavy inference serving. Interconnect efficiency directly sets system throughput and power in these workloads.
A programmable photonic network layer
The long-term goal is topology on demand: a software-defined photonic fabric that can reconfigure at packet cadence, respond to congestion and link health, and match connectivity to the workload in real time.
AI interconnects first. Broader photonics next
AI infrastructure is the first market. That's where this platform creates the clearest near-term advantage.
Over time, the same platform extends into adjacent domains including photonic compute primitives and quantum photonics.
Introducing Scale1
Our first photonic network switch
Scale1 is LightScale's first photonic network switch, built to turn the platform into a concrete interconnect product for AI infrastructure.
It is the first system expression of our Hyperlane architecture, with a focus on faster data movement, lower communication overhead, and more adaptive connectivity across bandwidth-heavy workloads.
Latency reduction
The switching path is designed to reconfigure at packet cadence, fast enough to change topology within a collective operation, not just between runs. The goal is to make connectivity changes cheap enough to use inside the data path, not as a slow control-plane action layered on top of it.
Programmable network layer
Topology, routing, and link assignment are exposed to the control plane. Operators can match fabric shape to workload patterns (training, inference serving, MoE routing) and feed congestion and link-health signals back into reconfiguration decisions in real time.
Hyperlane architecture
Scale1 is built on Hyperlane, a non-thermal photonic switching architecture, integrated at the device level rather than assembled from discrete optical components. It is designed to scale switch radix without the thermal load or per-port power overhead of conventional reconfigurable optics.
In development: Photon1
A photonic compute chip
Designed for AI training and inference with efficiency at scale.
Built on demonstrated photonics
Built on layered 2D semiconductor photonics, backed by published research and an active patent program, with a roadmap toward manufacturable photonic switching hardware.
Core functionality validated
Published work shows the platform's core photonic behavior in lab environments.
Process maturity
Key fabrication steps have already been demonstrated in the lab.
IP in progress
Patents have been submitted across the core architecture.
Built by researchers at the intersection of photonics and computing
LightScale brings together expertise across photonics, materials, fabrication, and large-scale systems, with backgrounds at MIT, Harvard, University of Toronto, Broadcom, Nvidia, Google and other leaders in photonics and AI infrastructure.
Leadership
Kabir Swain
Founder
JP Joseph
Materials Lead
Sijie Han
Architecture Lead
Manel Baradad
Software Lead
Tracey Ooi
Finance Lead
David Garrett
Advisory Board Member