Loading organizations...
§ Private Profile · Mountain View, CA, USA
Lightmatter is a technology company.
Lightmatter has raised $823.0M across 6 funding rounds.
Key people at Lightmatter.
Lightmatter has raised $823.0M in total across 6 funding rounds.
Lightmatter develops photonic chips and interconnect systems designed to accelerate artificial intelligence workloads. Their core products, Passage interconnects and Guide light engines, leverage silicon photonics to enable high-bandwidth, high-density communication within AI supercomputers, effectively overcoming the limitations of traditional electrical interconnects. These solutions facilitate rapid data transfer and processing for demanding AI models.
The company was founded in 2017 by Nicholas Harris, Darius Bunandar, and Thomas Graham. Their collective insight centered on the impending bottleneck that electrical data transfer would pose for the escalating computational demands of advanced AI. They recognized that a fundamental shift to photonic-based hardware was necessary to sustain the exponential growth of AI processing capabilities.
Lightmatter’s technology is adopted by enterprises pushing the boundaries of AI, particularly those involved in training large-scale, trillion-parameter models, multi-modal AI development, and inference at scale. The company envisions a future where its photonic computing and interconnect solutions unlock unprecedented performance and efficiency, enabling the next generation of intelligent systems by seamlessly merging light with computation.
Key people at Lightmatter.
Lightmatter has raised $823.0M in total across 6 funding rounds.
Lightmatter's investors include T. Rowe Price Associates, Breakthrough Energy Ventures, Craig Shapiro, Comcast Ventures, GV, Andrew Schoen, Prelude Ventures, Sequoia Capital, C2 Investment, Starbridge Venture Capital, Trajectory Ventures, Virginia Venture Partners.
Lightmatter has raised $823.0M across 6 funding rounds. Most recently, it raised $400.0M Series D in October 2024.
Lightmatter develops photonic chips and interconnects that use light instead of electrons for data processing and communication, targeting AI workloads and data centers.[1][2][3] Its core products—Passage (optical interconnect for high-bandwidth, low-power data transfer between chips) and Envise (photonic AI accelerator for efficient deep learning computations)—solve the limitations of traditional silicon computing by boosting speed, reducing energy use, and enabling massive scale-out.[1][4] Lightmatter serves hyperscalers, chipmakers, and AI firms facing exploding computational demands, with strong growth via $300M+ funding in 2023 (unicorn valuation), partnerships like Amkor for 3D-packaged chips in 2024, and demos with top tech companies.[1][2][4]
Lightmatter was co-founded in 2017 by MIT alumni Nicholas Harris (CEO, PhD '17 in photonics, ex-Micron Technology), Darius Bunandar (from MIT photonics lab), and Thomas Graham (MIT MBA).[2][4] Harris's doctoral thesis identified photonic solutions for quantum computing gaps, which he realized also applied to deep learning amid Moore's Law slowdown—prompting him to skip academia.[2][4] Early traction hit fast: won $100K in MIT's Entrepreneurship Competition in year one, secured a $4.8M U.S. grant in 2022 for electro-photonic tech in autonomous vehicles (with Harvard/Boston University), and expanded internationally by late 2024.[2] HQ in Boston (with Mountain View office), the team grew to 105 employees, blending photonics expertise with AI frameworks like PyTorch/TensorFlow.[3]
Lightmatter rides the AI compute explosion, where data centers hit physical limits from power-hungry GPUs amid decelerating Moore's Law (NVIDIA's CEO called it "dead" in 2022).[1][2][4] Timing is ideal: AI training demands hyperscale clusters, but copper interconnects cap bandwidth/energy efficiency—photonics unlocks "light-speed" scaling for next-gen platforms.[1][4][5] Market tailwinds include surging AI capex and sustainability mandates; Lightmatter influences by partnering with foundries/cloud providers, paving optical backbones that could cut data center emissions and democratize high-performance computing.[1][3][4]
Lightmatter's Passage is poised to underpin global AI GPU fleets, per CEO Harris, with mass deployment via chipmakers already underway.[1][4] Expect acceleration in 2026+ via expanded U.S./Canada teams, more 3D packaging ramps, and potential Idiom software integrations for full-stack photonic AI.[2][5] Trends like edge AI and exascale clusters will amplify its role, evolving from niche innovator to infrastructure standard—redefining compute efficiency as AI reshapes economies, much like it began with one MIT thesis challenging silicon's reign.[2][4]
