# Optics startup Quintessent secures $40M to revolutionize interconnects for AI

> Source: <https://www.sdxcentral.com/news/optics-startup-quintessent-secures-40m-to-revolutionize-interconnects-for-ai/>
> Published: 2026-08-24 15:09:11+00:00

[Quintessent](https://www.quintessent.com/), a startup developing optical interconnect solutions for AI data centers, has raised $40 million in a Series A funding round.

Canadian venture firm Cycle Capital led the oversubscribed round that included investment from Goldman Sachs XIG-Industry Ventures, Hina Liberty Capital, and M Ventures. Networking optics giant Ciena was a noteworthy name among the list of backers.

Quintessent demoed its dense wavelength division multiplexing (DWDM) comb laser technology at the OFC event earlier this year, claiming the technology can simplify optical connectivity for scaling AI clusters. The startup kicked off customer sampling and plans to use the cash to mature its product lines and ramp manufacturing.

“For the past several years, we have focused on building technologies designed to make optical connectivity fundamentally simpler to deploy and scale,” Quintessent CEO Alan Liu explained. “With today’s news, Quintessent is starting its transition from tech development to a product-focused company. We plan to use our investment to accelerate that transition and meet customer demand.”

Liu co-founded Quintessent alongside Brian Koch in 2019. The CEO was previously a lead engineer at Booz Allen Hamilton and a photonics consultant for the Defense Advanced Research Projects Agency (DARPA). Koch, who serves as the startup’s engineering VP, is an Intel and Juniper alum, notably having led the latter’s photonic device team.

They formed a startup working to reduce architectural complexity and improve reliability for optics. Their efforts focus on a DWDM link that generates eight precisely spaced wavelengths from a single laser, where traditional photonic deployments largely feature single individually tuned lasers that require additional wavelength-control units.

Quintessent’s architecture also looks to cut through the supply chain constraints of indium phosphide (InP), which is found in most conventional high-speed optics. Instead, the startup’s comb lasers use gallium arsenide (GaAs), an O-band quantum dot gain material obtained from more sourceable substrates that can be integrated on standard silicon photonics.

The startup claims its architecture enables up to a 40% reduction in data movement power compared to conventional narrow-and-fast architectures seen in deployment today. Quintessent goes as far as to claim its offering can be employed for simple bi-directional optical fiber communication with just one laser needed for each direction.

Quintessent is also looking to develop other “critically needed optical components,” including semiconductor optical amplifiers (SOAs), which are used to boost weak signals in fiber-optic data links, as well as an optical engine for pluggable interconnect modules for AI data centers.

“AI continues to accelerate data center demand, placing significant pressure on power systems. With its novel quantum dot material and disruptive comb laser design, Quintessent will be instrumental to scaling up data center performance while also reducing energy consumption,” Cycle Capital Founder and Managing Partner Andrée-Lise Méthot added.

AI demand is set to cause silicon photonic revenues to [surge](https://www.sdxcentral.com/news/ai-demand-to-fuel-23x-silicon-photonic-growth-by-2030/) over the next several years, according to recent market projections. Quintessent finds itself in a market buoyant with emerging players looking to capture that optics interest, following recent raises from [Lumilens](https://www.sdxcentral.com/news/ex-aruba-cto-unveils-optical-startup-armed-with-900m-to-scale-ai-interconnectivity/), [Olix](https://www.sdxcentral.com/news/sdn-pioneer-joins-33b-startup-building-optical-interconnects-for-ai-infrastructure/), and [OptQC](https://www.sdxcentral.com/news/ntt-invests-in-optqc-to-accelerate-optical-quantum-computing/), among a host of others.
