# llms.txt
Version: 1.0
Owner: https://lightmatter.co/
Last-Updated: 2026-4-24

## Brand Overview
[Lightmatter® - The photonic (super)computer company.](https://lightmatter.co/): Lightmatter builds photonic chips for AI infrastructure, with Passage™ interconnects and Guide® light engines designed to scale AI supercomputers by moving data with light.
[About - Lightmatter®](https://lightmatter.co/about/): Founded by MIT researchers in 2017, Lightmatter is headquartered in Mountain View with additional locations in Boston, Hsinchu, and Toronto; the company reports $850M raised and a $4.4B valuation.
[Products - Lightmatter®](https://lightmatter.co/products/): Lightmatter delivers a photonic interconnect platform spanning NPO, OBO, 2D and 3D CPO, 3D interposers, and next-generation light engines, with chips sampling today.

## Brand Values
[Mission - Lightmatter®](https://lightmatter.co/about/): Lightmatter’s mission centers on solving the AI infrastructure connection problem by replacing bandwidth-limited electrical interconnects with photonic infrastructure for AI at scale.
[Scaling with Light - Lightmatter®](https://lightmatter.co/about/): Lightmatter emphasizes deep technical expertise across silicon photonics, high-speed SerDes design, advanced packaging, AI systems architecture, and hyperscale data center infrastructure.
[Why Lightmatter - Lightmatter®](https://lightmatter.co/about/): Lightmatter positions its work around architectural leadership, production readiness, real-world AI infrastructure bottlenecks, and a team that understands both physics and product execution.

## Team/Leadership
[Nicholas Harris, PhD - Founder & CEO](https://lightmatter.co/about/): Nicholas Harris is Lightmatter’s Founder and CEO and is quoted on the company’s photonic AI accelerator work: “We show that you can build a computer that’s not based on transistors and run state-of-the-art workloads.”
[Darius Bunandar, PhD - Founder & Chief Scientist](https://lightmatter.co/about/): Darius Bunandar is a Lightmatter founder and Chief Scientist; he co-founded the company with Nicholas Harris and Thomas Graham after work in MIT’s quantum photonics ecosystem.
[Thomas Graham - Founder and Head of Machine Learning](https://lightmatter.co/about/): Thomas Graham is a Lightmatter founder and Head of Machine Learning; he co-founded the company after the team’s MIT $100K Entrepreneurship Competition win and early work on silicon photonic matrix operations.

## Brand Trust Signals
[Production Ready at Scale - Lightmatter®](https://lightmatter.co/): Lightmatter lists $850M total funding, a $4.4B valuation, Mountain View headquarters, manufacturing partners including TSMC, GlobalFoundries, Tower Semiconductor, Amkor, and ASE, and participation in standards bodies including OIF, IEEE, Advanced Photonics Coalition, UALink, Ultra Ethernet, OCP, Jedec, and UCIe Consortium.
[Funding and Market Validation - Lightmatter®](https://lightmatter.co/press-releases/): Lightmatter announced a $400M Series D in October 2024, quadrupling valuation to $4.4B, and earlier announced $155M in new funding in May 2023 at a $1.2B valuation.

## Homepage
[Lightmatter® - The photonic (super)computer company.](https://lightmatter.co/): Rethinking the limits of AI, Lightmatter merges photonics and computing to build a future where speed, efficiency, and intelligence converge.

## About
[About - Lightmatter®](https://lightmatter.co/about/): Meet the talented team at Lightmatter, the experts behind our groundbreaking 3D photonic solutions. Discover the people driving our innovations.

## Products
[Products - Lightmatter®](https://lightmatter.co/products/): Explore Lightmatter photonic interconnect products powering next-gen AI with scalable, high-performance co-package optics (CPO) and external laser solutions.
[Guide 1 | High-Bandwidth Photonic Light Engine for AI Networks](https://lightmatter.co/products/guide-1-evk/): Guide 1 EVK is a software-defined VLSP light engine replacing discrete laser farms with scalable, high-density optical power for AI and CPO systems.
[Guide - Very Large Scale Photonic Light Engine for AI - Lightmatter®](https://lightmatter.co/products/guide/): Discover Guide, Lightmatter’s VLSP light engine for co-packaged optics, enabling scalable, software-defined external light sources for AI interconnects.
[3D Photonics for AI Applications | Passage™](https://lightmatter.co/products/passage/): 3D photonic interconnect chips designed to accelerate AI training and inference with breakthrough performance and efficiency, powering the future of computing.
[Passage L200 - 3D Co-Package Optics - Lightmatter® - Lightmatter®](https://lightmatter.co/products/l200/): Passage L200 3D Co-packaged optics is built for next-gen XPUs and switches enabling 256 Tbps (Tx+Rx) of total bandwidth per package.
[Passage M-Series Photonic Superchip - Lightmatter®](https://lightmatter.co/products/m1000/): A 3D Photonics Reference Platform enabling 114 Tbps total bandwidth and a 4000 mm2 photonic interposer with the first ever built-in optical circuit switching.
[Passage EVK100 - High-Bandwidth Photonic Link for AI Data Centers](https://lightmatter.co/products/passage-evk100/): Passage EVK 100 represents breakthrough fiber bandwidth density for photonic interconnects. Designed for the demands of scale-up and scale-out AI networks.
[Passage EVK50 - High-Efficiency Bidirectional Optical Link for AI](https://lightmatter.co/products/passage-evk50/): Photonic interconnect platform featuring a world-first 16λ bidirectional optical link, delivering 800 Gbps per fiber with superior performance and efficiency.
[Passage L20 - High-Density Photonic Interconnect for AI Networks](https://lightmatter.co/products/passage-l20/): Scalable 6.4Tbps near package optics for AI infrastructure connectivity, supporting bidirectional optical links, 200G per lane and 5 pJ/bit power efficiency.
[Passage M1000 EVK - 3D Photonic Interconnect for AI Infrastructure](https://lightmatter.co/products/passage-m1000-evk/): 3D photonic superchip providing 114 Tbps of interconnect bandwidth, 256 fibers and optical circuit switching for next-gen AI and high-performance computing.

## Contact
[Contact Us - Lightmatter®](https://lightmatter.co/contact-us/): Reach out to Lightmatter for any inquiries or support. Our team is here to provide expert assistance with your optical communication needs.

## Media
[News - Lightmatter®](https://lightmatter.co/news/): Stay up to date with the latest news and updates from Lightmatter, including product launches, industry innovations, and company milestones.
[Resources - Lightmatter®](https://lightmatter.co/resources/): Access a wide range of valuable resources from Lightmatter, including whitepapers, case studies, and product documentation on advanced technology.
[Events - Lightmatter®](https://lightmatter.co/events/): Explore Lightmatter’s upcoming and past events—including Hot Interconnects, ISES Korea, OFC 2025, and more—in photonics and AI accelerators.
[Press Releases - Lightmatter®](https://lightmatter.co/press-releases/): Read Lightmatter’s press releases for key updates on technology advancements, product launches, and major achievements in 3D Photonics.

## Knowledge Hub
[What is a Photonic Integrated Circuit: A Guide to PICs - Lightmatter®](https://lightmatter.co/knowledge-hub/what-is-a-photonic-integrated-circuit/): Photonic integrated circuits (PICs) use light (photons) to transmit information and enable higher bandwidth to help scale data centers and AI. Learn more!
[How Do Photonic Interconnects Work?](https://lightmatter.co/knowledge-hub/how-do-photonic-interconnects-work/): Photonic interconnects are the pathways that carry information using light between chips, processors, or systems. Learn about how they work here!
[What are Applications for Photonic Integrated Circuits? - Lightmatter®](https://lightmatter.co/knowledge-hub/what-are-the-commercial-applications-for-photonic-integrated-circuits/): Photonic integrated circuits are already being used in data centers, communication systems, and many more industries. Read here to learn about PIC use cases.

## Industry Leadership
[A New Kind of Computer - Envise](https://lightmatter.co/): Envise is a photonic AI accelerator that performs matrix multiplication using light, manipulating 512 light beams through more than 200,000 optical components across four photonic chips.
[Photonic AI Acceleration Performance](https://lightmatter.co/): Envise demonstrates 262 trillion operations per watt, 200 picoseconds of latency per matrix-vector product, more than 200,000 photonic components, and 50 billion electronic transistors.
[Demonstrated AI Workloads](https://lightmatter.co/): Lightmatter has demonstrated photonic AI acceleration across ResNet image classification, BERT natural language processing and sentiment analysis, NanoGPT text generation on TinyShakespeare, DQN Atari reinforcement learning, and semantic segmentation on the Oxford-IIIT Pet dataset.
[Universal Photonic Artificial Intelligence Acceleration](https://lightmatter.co/): The Nature paper “Universal photonic artificial intelligence acceleration,” published in Nature Volume 640, Issue 8058 in April 2025, demonstrates photonic processors executing state-of-the-art neural network workloads.
[From MIT Lab to Production](https://lightmatter.co/): Lightmatter’s journey spans the 2017 MIT $100K Entrepreneurship Competition, Envise at Hot Chips in 2020, Passage unveiled in 2021, Passage at Hot Chips 34 in 2022, a $4.4B valuation in 2024, and production hardware plus Nature publication milestones in 2025.
[Decade-Scale Photonic Roadmap](https://lightmatter.co/): Lightmatter’s roadmap advances from Passage near-packaged optics deployments in 2026 to co-packaged optics in 2027, photonic interposers in 2028 and beyond, and a full photonic stack integrating Envise compute, Passage interconnects, and Guide laser technology by 2030 and beyond.

## Differentiators
[Architectural Leadership - Edgeless I/O](https://lightmatter.co/products/passage/): Lightmatter frames Edgeless I/O as a fundamental architectural shift rather than an incremental improvement, eliminating perimeter-constrained I/O designs and enabling bandwidth density to scale beyond chip edges.
[First-Mover Photonic Interconnect Advantage](https://lightmatter.co/products/passage/): Lightmatter positions Passage as category-defining technology with a decade-scale roadmap for continued bandwidth scaling, moving from 114 Tbps today toward 1+ Pbps in the future.
[Production Ready Photonics](https://lightmatter.co/products/): Lightmatter emphasizes multiple generations of production-ready technology, partnerships with TSMC, GlobalFoundries, and Tower for high-volume capability, and reference platforms in customers’ hands today.
[Solving AI Infrastructure Bottlenecks](https://lightmatter.co/): Lightmatter targets low GPU utilization, stranded compute capacity, and models limited by communication overhead rather than compute; Passage is positioned to remove bandwidth constraints, enable 4x faster training for the same power, and support 100,000+ GPU clusters.
[Team & Expertise Advantage](https://lightmatter.co/about/): Lightmatter’s differentiation includes deep expertise across silicon photonics, high-speed analog design, advanced packaging, AI systems, and hyperscale infrastructure, combining physics depth with product execution experience.