Lightmatter

Lightmatter builds photonic processors and optical interconnects to accelerate AI compute and reduce data-center power consumption. Raised $400M+ including Series D in 2024; Passage interconnect fabric announced. Partnerships with GlobalFoundries and others.

Lightmatter's silicon photonics approach addresses the power and bandwidth walls limiting electronic AI accelerators.

lightmatter.co Photonics & InterconnectsUSAprivate

At a glance

Total raised

$667M

last round series d

Headcount

—

no public estimate on record

Headquarters

Boston

USA

Founded

2017

9 years active

Status

Private

Photonics & Interconnects

Latest valuation

$4.4B

Series D · 2024

For subscribers

Subscribe to see the full profile

Including Funding rounds, Bull / Bear thesis, Stock + earnings, Roster changes, Patents, News, and Open roles.

Founding

50% off

/month

91 of 100 spots left

Full

$10

/month

Available once all 100 Founding Member spots are claimed.

Get full access

Already subscribed? Sign in →

Funding history

Total raised

$667M

Rounds tracked

Latest stage

Series D

Latest valuation

$4.4B

Date	Stage	Amount	Valuation	Lead investors
Oct 16, 2024	Series D	$400M	$4.4B	T. Rowe Price
May 1, 2023	Series C	$154M	—	GV
May 6, 2021	Series B	$80M	—	Viking Global Investors
Jan 1, 2019	Series A	$33M	—	GV

How they make money

Hardware + recurring servicesPattern

Lightmatter runs a hardware-plus-recurring-services model: customers purchase photonic processors and optical interconnect hardware for AI compute clusters, then pay ongoing fees for support, maintenance, and firmware updates tied to the deployed systems. The company sells silicon-photonic chips and the Passage interconnect fabric to hyperscale data-center operators, AI labs, and cloud providers who need to move data between accelerators faster and with less power than copper or traditional optics. Revenue is recognized on hardware delivery plus periodic service contracts for each installation. Hardware gross margins in semiconductor fabrication vary with process yields and volume, though the recurring service layer—support contracts, software tooling for photonic circuit management, and integration engineering—carries higher incremental margins once systems are deployed. Manufacturing is outsourced to foundry partners such as GlobalFoundries, so capital expenditure for fabrication is largely avoided, but the company carries R&D and wafer-design costs upfront. The model benefits as customers scale clusters: each additional node requires more photonic interconnects, pulling through both hardware sales and service renewals without proportional increases in sales cost.

Generated by deepseek-v4-flash · June 2, 2026Analytical summary · not investment advice

Founder backgrounds

Nick Harris

Co-founder & CEO

Holds a Ph.D. in electrical engineering and computer science from MIT, where his doctoral thesis — 'Programmable Nanophotonics for Quantum Information Processing and Artificial Intelligence' — laid the research foundation for Lightmatter. Was an NSF Fellow and Intelligence Community Postdoctoral Fellow at MIT. Previously worked as an R&D engineer at Micron Technology on DRAM and NAND circuits. Founded Lightmatter in September 2017 to commercialize silicon photonics for AI compute, developing the Passage interconnect fabric and Envise photonic inference processor. Holds 87 or more patents and is an Optica Fellow.

Darius Bunandar

Co-founder & Chief Scientist

A physicist and photonics researcher who co-founded Lightmatter in 2017 based on silicon photonics research conducted at MIT. Serves as Chief Scientist, leading the scientific and research direction of the company's photonic computing platform. His academic background in quantum optics and silicon photonics integration informs Lightmatter's approach to using light rather than electrons for chip-to-chip and on-chip data movement, targeting the bandwidth and energy bottlenecks that limit large-scale AI training and inference systems in data centers.

Thomas Graham

Co-founder (former COO & CFO)

Co-founded Lightmatter in 2017 alongside Nick Harris and Darius Bunandar, serving in both COO and CFO capacities during the company's early growth phase. Helped build the operational and financial infrastructure as the company scaled from its MIT photonics research roots into a venture-backed photonic computing startup. Lightmatter raised over $340 million in total funding, including a $154 million Series C, to develop its Passage photonic interconnect and Envise inference processor targeting AI data center applications.

AI bull / bear

Bull case

Lightmatter has raised $400M+ through a Series D in 2024, signaling deep institutional conviction that photonic interconnects are the answer to AI's power and bandwidth crisis. Its Passage interconnect fabric targets the highest pain point in scaling: moving data between exponentially growing clusters of accelerators. The partnership with GlobalFoundries provides a credible manufacturing path that most photonics startups lack. As Nvidia's H100/B200 systems push 700W+ per GPU and data-center power draw becomes a binding constraint, any technology that decouples communication from electronic loss becomes indispensable. If Passage delivers on its latency and energy-per-bit promises at rack scale, Lightmatter owns a layer of the AI stack that electronic interconnects cannot replicate.

Bear case

Photonic computing has a decades-long track record of promising breakthroughs and delivering only niche laboratory deployments. Lightmatter has not disclosed meaningful customer revenue, and its $400M+ in funding buys it time but not traction against incumbents that already own the data-center fabric. Nvidia's NVLink and InfiniBand advance on a yearly cadence with full software-stack lock-in, making optical interconnects a solution in search of a switching problem that Nvidia may preempt. Manufacturing silicon photonics at yield parity with mature CMOS nodes is unproven at the volumes AI clusters demand. Meanwhile, Broadcom and Cisco are pushing co-packaged optics directly into hyperscaler procurement pipelines. Lightmatter risks being a technology marvel that hyperscalers sample but never deploy at scale.

Generated by deepseek-v4-flash · June 2, 2026Refreshes quarterly · see disclosures

In the news

Lightmatter · last 30 days

No articles ingested yet for Lightmatter. Once the hourly news pipeline is live, every article the classifier tags as mentioning this company appears here with its one-line AI summary and sentiment.

Watch

Leadership

Nick Harris

Co-founder & CEO & Chief Executive Officer

Darius Bunandar

Co-founder & Chief Scientist

—

Thomas Graham

Co-founder (former COO & CFO)

—

Products

Passage M1000 Photonic Superchip
Lightmatter's flagship 3D photonic interconnect, described as the world's first photonic interposer: a wafer-scale active substrate that stacks compute chiplets on top while routing terabits of data optically across its surface. The M1000 delivers multi-terabit-per-second per-port bandwidth to connect GPUs, TPUs, and switches in the largest AI training clusters, attacking the interconnect bandwidth wall that limits how big AI systems can scale.
Passage L200 Co-Packaged Optics
A chiplet-style co-packaged optics product that brings optical I/O directly next to an accelerator or switch ASIC, replacing power-hungry electrical SerDes with light. The L200 dramatically raises the bandwidth that can escape a package while cutting energy per bit, enabling denser, more efficient scale-up and scale-out fabrics for AI data centers.
Envise & Idiom Software
Envise is Lightmatter's photonic compute engine that performs neural-network matrix operations using light for high throughput at low power, paired with Idiom, a software stack that lets standard ML models (PyTorch/TensorFlow) compile down to the photonic hardware. Together they target energy-efficient AI inference, complementing Lightmatter's photonic interconnect roadmap.

Most recent patents

31 patents on file, but none with both an extractable figure and an abstract on Google Patents yet.

Compare Lightmatter with

vs Photonics & Interconnects
Lightmatter vs Ayar Labs