co-packaged-optics
(5)
Ayar Labs
Ayar Labs is a San Jose-based semiconductor company building optical input/output (I/O) technology that moves data between chips using light instead of copper. Its silicon-photonics chiplets - the TeraPHY optical engine and the SuperNova multi-wavelength light source - are designed to be co-packaged directly with GPUs and other processors, letting AI and high-performance-computing systems scale across thousands of accelerators with far higher bandwidth and far lower power than electrical interconnects. Spun out of a DARPA-funded university research effort, the company raised a $500M Series E in 2026 with backing from NVIDIA and AMD.
SCINTIL Photonics
SCINTIL Photonics is a Grenoble-based, fabless silicon photonics company building the world's first single-chip, DWDM-native laser source for AI data centers. Spun out of CEA-Leti and founded by Sylvie Menezo, its proprietary SHIP (Scintil Heterogeneous Integrated Photonics) process integrates lasers, modulators and detectors onto one manufacturable chip - powering its LEAF Light light engine for co-packaged optics that promises roughly one-sixth the power consumption of conventional pluggable optics. The company raised a $58M Series B in September 2025 with participation from NVIDIA.
Lightelligence
Lightelligence builds photonic and optoelectronic hybrid computing hardware that moves and processes AI data with light instead of electrons. Spun out of MIT in 2017 on the back of a landmark Nature Photonics paper, the company designs silicon-photonics chips, optical interconnects, and AI accelerators aimed at the bandwidth and energy bottlenecks of modern data centers. In April 2026 it listed in Hong Kong as the world's first AI silicon-photonics chip stock, jumping roughly 400% on debut.
Avicena Tech
Avicena Tech builds LightBundle, a microLED-based optical interconnect that moves data between chips at terabit-per-second densities while sipping sub-picojoules of energy per bit. The Sunnyvale company is betting that the GPU clusters powering AI need a new physical layer - one that copper can't deliver and traditional silicon photonics can't match on power.
Lightmatter
Lightmatter is building a photonic supercomputer. The Mountain View company uses light, not electrons, to move data between AI chips - tackling the bandwidth and energy wall that's about to crash into the next generation of data centers. Its Passage interconnect and Envise processor aim to connect millions of chips at the speed of light.