Ayar Labs Extends Series E to $650 Million to Scale Optical AI Chip Clusters

The extension follows the company's original $500 million Series E, disclosed in March, which was led by Neuberger Berman and included Nvidia Corp., Advanced Micro Devices Inc. and MediaTek Inc. That round had brought Ayar Labs' total raised to $870 million. With the extension, the startup has raised just over $1 billion in outside funding.

Ayar Labs' optical technology aims to replace copper interconnects in AI data centers. Copper has three main limitations: signals degrade over distance, the cables consume significant power when thousands of GPUs are connected, and they can overheat when linked to multiple high-performance processors, causing workloads to break down. Ayar Labs says optical interconnects offer greater bandwidth, better performance, higher power efficiency and lower latency, potentially allowing much larger GPU clusters across data centers.

The company's main product, TeraPHY, substitutes copper with optical interconnects. Ayar Labs has also developed co-packaged optics technology that places its TeraPHY Optical Engine directly in a server's ASIC socket alongside the GPU. Mark Wade, co-founder and CEO, said copper interconnect is becoming the limiting factor for AI scale-up and that the funding will help Ayar Labs move faster on manufacturing-ready co-packaged optics at scale. He said the company is deepening work across the foundry and packaging ecosystem and expanding engineering capacity for customer programs worldwide.

Ayar Labs was founded in 2015 and did not raise significant outside capital until five years later, when it received $35 million. The AI boom has increased urgency around optical interconnect development, and some of its largest backers are chipmakers that could benefit from larger chip clusters. Wade told Reuters that the company hopes to bring its technology to the mass market by early 2028. The Series E money has already helped pay for a new chip design center in Bengaluru, India, and the extension will help Ayar Labs prepare its products.

Wade said timing is tight because Ayar Labs supplies optical chips that go into customers' products. If customers plan to ramp in the 2028-2029 period, Ayar Labs must have all of its offerings qualified for volume production by the end of 2027, he said.

Ayar Labs expects help from Wiwynn Corp., a Taiwanese data center infrastructure company that also invested in it. Ayar Labs said Wiwynn made a strategic investment earlier this year, though it is not clear whether that is the source of the Series E extension. The companies have worked together to deliver optically connected, rack-scale AI systems for hyperscale workloads. The systems use SuperNova remote light source technology integrated directly into Wiwynn's rack-level server architecture.

Wiwynn President and CEO William Lin said co-packaged optics is a foundational technology for next-generation AI and cloud data centers. He said the collaboration and investment with Ayar Labs offers a path to architectures that reduce the power and complexity limits of copper interconnects and support rack-scale deployment for the next wave of hyperscale AI data centers.

Editor's Summary

Ayar Labs extended its Series E round to $650 million after raising $150 million from an unnamed investor, bringing total outside funding to just over $1 billion. The company is developing optical interconnects and co-packaged optics meant to replace copper links and support larger AI chip clusters. It aims for mass-market availability by early 2028 and needs volume-production qualification by the end of 2027.