Ayar Labs, the San Jose, California-based developer of optical chip interconnects for AI data centers, has added $150 million to its Series E, bringing the round to $650 million. The extension lifts the company's total funding to just over $1 billion.

Neuberger Berman, which led the Series E, returned alongside strategic backers including NVIDIA, AMD, MediaTek and server maker Wiwynn.

Light Instead of Copper

Inside today's AI clusters, GPUs talk to each other over copper wiring. As clusters grow to tens of thousands of chips, those electrical links run into hard limits on distance, bandwidth, and power consumption.

Ayar Labs replaces them with light. Its TeraPHY optical engine uses co-packaged optics, placing optical interconnects directly alongside processors so data moves between chips as photons rather than electrons. The result, the company says, is more bandwidth, lower latency, and better power efficiency, enabling larger GPU clusters.

"Copper interconnect is becoming the limiting factor for AI scale-up," said co-founder and CEO Mark Wade. "This funding allows us to move faster on manufacturing-ready co-packaged optics at scale."

From Lab to Volume Production

Ayar Labs was founded in 2015 and raised its first significant round, $35 million, in 2020. After closing a $500 million Series E earlier this year, the company is now focused on moving its technology into high-volume manufacturing, with a mass-market launch planned by early 2028.

The new capital will fund manufacturing-ready co-packaged optics, expand its foundry and packaging ecosystem, and grow engineering capacity for global customer programs. That includes a new chip design center in Bengaluru, India, SiliconANGLE reported.

Strategic Backers

The investor list reads like a map of the AI hardware supply chain. NVIDIA and AMD design the accelerators that would use Ayar's interconnects, MediaTek builds custom silicon for hyperscalers, and Wiwynn assembles the servers.

"Co-packaged optics is a foundational technology for the next generation of AI and cloud data centers," said Wiwynn CEO William Lin.

Why It Matters

As AI models grow, the bottleneck is shifting from individual chip power to how fast those chips can share data. Optical interconnects are widely viewed as the next step, and the race to get co-packaged optics into production has drawn in both startups and chip giants.

With more than $1 billion raised and backing from the companies that would ultimately deploy its technology, Ayar Labs is one of the best-capitalized independent players in that race. The next milestone is proving it can ship at the volumes hyperscalers demand.

Source

SiliconANGLE — "Ayar Labs bags $150M in additional Series E funding to help make bigger AI chip clusters"