[REVISION HISTORY]

Advancements in silicon photonics for AI infrastructure

1 recorded revision of this situation.

What changed

2026-10-19 02:16 UTC → 2026-10-21 07:31 UTC · added ~~removed~~

The technology sector is seeing increased development in silicon photonics and integrated optics to support next-generation artificial intelligence and high-performance computing. Coherent introduced its PhotonLink platform, an integrated optics solution designed for AI data centers. The platform utilizes technologies such as InP lasers and silicon photonics to support co-packaged and near-packaged optics. The company reported multiple engagements with anchor customers and expects revenue to begin ramping in late 2026. Following this, Intel positioned silicon photonics as a critical solution for overcoming the limitations of traditional copper interconnects, specifically regarding bandwidth density and energy consumption. Intel highlighted its capacity to provide 300mm Indium Phosphide silicon photonics technology at scale to support near-package and co-packaged optics. The company aims to drive industry standards for external light sources and reduce energy consumption to below 5 picojoules per bit. At the 2026 Open Compute Project (OCP) Global Summit, industry leaders emphasized a shift in AI infrastructure toward high-density, integrated systems. Nvidia introduced design contributions for its Vera Rubin platform via the MGX 1.2 specification, featuring rack-scale liquid cooling and power densities up to 250 kW per rack. This transition positions liquid cooling as a critical requirement for modern compute trays. Intel reaffirmed its focus on silicon photonics to address bandwidth limits, stating it is currently the only foundry capable of large-scale industrial production of 300 mm silicon photonics solutions integrating indium phosphide (InP).

Versions

Select a version to see what changed compared to the one before it.