Ayar Labs announced an additional $150 million on September 10 for the optical components that connect AI computing hardware. Its customers are designing chips and systems that need to move large volumes of data between processors. The financing backs a transition from technology development toward manufacturing those components at greater scale.
The core product is TeraPHY, an optical engine designed to sit alongside computing silicon within a chip package. Instead of carrying all data over copper connections, it sends information through optical links. For a system designer trying to make processors in multiple racks work together, the attraction is a connection that can extend beyond a single board or enclosure.
The engine works with an external light source, including Ayar Labs’ SuperNova product. Separating that source from the engine is intended to help with heat management and replacement in the field. The company presents the two as components for integration into a larger system, not a complete AI server that replaces a customer’s processors.
Integration is a substantial part of the business. The product materials describe compatibility with established chip-packaging architectures, electrical interfaces and detachable fiber connectors. That gives semiconductor and infrastructure partners a route to incorporate the technology into their designs. Published performance advantages remain vendor claims, while reliability testing and manufacturable packaging are practical requirements for customers deciding to deploy it.
The new financing brings primary capital raised during 2026 to $650 million, according to the company; only the additional $150 million is newly announced here. FinSMEs also reported the raise. Ayar separately disclosed an investment by Wiwynn made earlier in the year, so that disclosure should not be mistaken for another new round. The proceeds support development, validation, manufacturing readiness and a Bengaluru engineering center.
The opportunity is to make optical connections routine parts of AI hardware rather than special demonstrations. That requires coordinating chips, light sources, packaging and system assembly across several suppliers. Ayar’s next step is therefore as much about repeatable production and customer integration as the speed of an individual optical link.