ARIA to Test Lumai Optical AI Inference Servers

Nov 3, 2026
ARIA will deploy Lumai Iris optical inference servers in its Scaling Inference Lab, where researchers will test the technology with AI workloads in a commercial data center.

Advanced Research and Invention Agency (ARIA) will deploy optical inference servers from Lumai in its Scaling Inference Lab, Lumai stated in an announcement. The lab, operated by CommonAI CIC, will test the servers with AI workloads in a commercial data center using standard power, cooling and rack systems.

Lumai Iris servers perform AI calculations using beams of light and passive optical components. The system targets the compute intensive prefill phase of large language model inference and works alongside silicon hardware that handles decoding.

Lumai says Iris can provide at least 16 times greater energy efficiency per inference than systems using only silicon. ARIA will evaluate those claims, including the potential to reduce power consumption by 90% and total ownership costs to about one tenth of conventional GPU systems.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Silicon Brief or Daily AI Brief.

Also, consider following us on social media:

AI Chips & Datacenters AI Brief AI Brief (X)
Free newsletter

Silicon Brief

Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.

Guide
AI interconnect and networking startups

AI interconnect and networking startups

29 private startups developing the chips, optical switches and memory fabrics that move data through AI infrastructure.

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more

You may also like

Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.