Lightmatter Brand Guidelines: The MIT Photonics Spin-Out Moving AI Data With Light

Three MIT alumni and the $100K prize.

Lightmatter grew out of a photonics laboratory at the Massachusetts Institute of Technology. The company says it was “founded by MIT researchers in 2017.” MIT News reported in March 2024 that its three founders were MIT alumni: Nicholas Harris, who completed his PhD in 2017; Darius Bunandar, who earned his PhD in 2018 in the same laboratory; and Thomas Graham, an MBA graduate of 2018. Harris remains chief executive, Bunandar chief scientist and Graham head of machine learning, the company’s website says.

Harris had worked at the memory maker Micron Technology before MIT, where he did his doctorate on photonic quantum computing under the electrical engineer Dirk Englund, building silicon chips that send and process information with light rather than electricity, MIT News wrote. He told the publication that the team realised the systems it had built for photonic quantum computing could also be used for deep learning, the AI technique he recalled students crowding into lectures to learn. Englund is now listed as a technical advisor to the company.

The team’s public debut came in May 2017, when it won the Robert P. Goldberg grand prize at the MIT $100K Entrepreneurship Competition, then in its 28th year, with a pitch for chips that compute using light. MIT News reported on 18 May 2017 that Harris, then a PhD student, told the audience in Kresge Auditorium: “We’re bringing the next step of artificial intelligence to light.” MIT News said in 2024 that Harris had planned to become a professor but concluded that a startup could attract more funding and move faster.

Envise, Passage and Guide.

Lightmatter’s first two products were named Envise and Passage. MIT News described Envise in 2024 as a chip that leaves to electronics what they do well, such as memory, and uses light for the large matrix multiplications at the heart of deep-learning models. Passage is an interconnect that links processors with light, much as fibre-optic cables carry data over long distances, and lets chips as large as whole wafers act as a single processor. Harris told the publication that because the technology runs on silicon, it can be produced in existing semiconductor fabrication plants without large changes in process.

The company’s current website leads with interconnects and lasers. On 31 March 2025 it announced the Passage M1000, which it calls a “3D Photonic Superchip”: a reference platform of more than 4,000 square millimetres in which an active photonic interposer, a layer beneath the stacked chips, carries optical signals across its whole surface rather than only at the edges. It is built on GlobalFoundries’ GF Fotonix silicon photonics platform, and Lightmatter worked with GlobalFoundries and the packaging company Amkor on production readiness, the company said; the trade publication Evertiq reported the same details on 18 April 2025. The same day Lightmatter announced the Passage L200, a co-packaged optics product that connects to processors through the UCIe die-to-die interface standard.

Later announcements added a laser line. The Guide light engine is built on what the company calls Very Large Scale Photonics, which it says moves laser manufacturing from manual assembly lines towards foundry production. Lightmatter has also announced a partnership with Global Unichip Corp., the chip design services company in Hsinchu, Taiwan, to bring Passage co-packaged optics to market. On its About page, the company says it builds “interconnects, lasers, and eventually compute itself.”

How AI chip companies present themselves.

AI chip companies sell something most of their customers will never see. The processor sits inside a server, the server sits in a rack, and the rack sits in a data hall far from the people who use the models running on it. So the industry’s visual language has grown up around making that hidden hardware visible: a whole silicon wafer, as wide as a dinner plate, held up on a keynote stage; a single chip pinched between finger and thumb for the cameras; a die shot recoloured in neon false colour so that its blocks of memory and compute read like a city map at night.

The second set of images is the data centre. Press photographs tend to show long rows of racks in dark halls, lit by the glow of status lights and cold aisles, with cabling arranged as neatly as a diagram. These pictures stand in for scale, which is what buyers of AI compute are being asked to imagine, and they appear again and again in launch announcements, partner news and investor material.

Keynote slides do much of the remaining work. Model names, token counts and benchmark charts appear beside the chip’s own product name, and the companies’ identities are usually kept spare to let those numbers and images lead: black or white grounds, a single accent colour, a sans-serif wordmark and very little ornament. Several firms have adopted coined product categories, names for their processors that set them apart from the graphics processor.

The shop window is the technical conference. Hot Chips, NeurIPS and the AI Infra Summit are among the events where architectures are presented, demonstrations run and booths compete for engineers’ attention. This guide looks at how each company in the group presents itself, and keeps to what can be checked about who founded it, who owns it and what it makes.

Boston roots, a Mountain View headquarters.

Lightmatter is privately held and funded by venture investors. Its base has been described in two ways: Evertiq called it a Boston-headquartered startup in April 2025, while the company’s own website now lists Mountain View, California, as its headquarters, with further offices in Boston, Hsinchu and Toronto. Harris and Bunandar sit on its board alongside outside directors, the website says.

On 3 June 2026, in Taipei, Lightmatter announced that it had joined Nvidia’s NVLink Fusion ecosystem. It said it would deliver co-packaged and near-packaged optics compatible with Nvidia’s optical and SerDes technologies, so that customers’ semi-custom processors could connect to Nvidia switch silicon. In its press releases the company describes itself as “the leader in photonic (super)computing,” with the brackets its own.

The name joins two plain words, light and matter, written as one word with a capital L. The company marks it with the registered-trademark sign on its website, and product names such as Passage and Guide carry a trademark sign. The website opens with a short line: “We’re making sure computers keep getting better. With light.” Its explanations lean on one image, the “shoreline,” its word for the edge of a chip, where electrical bandwidth runs out, a limit the company says light does not share.

For a design guide, Lightmatter shows how a company built on hard physics turns its idea, moving data with light instead of electrons alone, into short, plain words and a single image. Compare Cerebras, another AI chip company in this series, and Nvidia, whose NVLink Fusion ecosystem Lightmatter joined in 2026. For this guide we use near-black, a warm amber and a soft white with a sans, an unofficial interpretation; no product or logo is reproduced.

Keep Lightmatter’s identity on hand — save it to a board in the Made Good inspiration library.