Analysis
Ayar Labs has raised $150 million in additional funding, SiliconANGLE reported, extending a year of financings that now stack up as follows:
- March 2026 -- $500M Series E closed, valuing the company at $3.75B.
- Secondary sale -- $225M transaction, pricing the company above $5B, a 33% markup.
- This week's tranche -- $150M more, bringing 2026's cumulative total to $650M.
Ayar Labs builds optical interconnect technology -- chips that link AI accelerators to fiber-optic cables instead of copper, cutting the latency and power draw that bottleneck data transfer inside large GPU clusters. As AI training runs stretch across tens of thousands of chips, the connections between them matter almost as much as the chips themselves, which is why Ayar Labs counts strategic backers across the entire AI hardware stack rather than just financial investors:
“- March 2026 -- $500M Series E closed, valuing the company at $3.75B.”
- Nvidia and AMD -- the two dominant AI accelerator makers, both betting Ayar Labs' co-packaged optics (CPO) becomes standard for scaling clusters beyond a single rack.
- Intel and MediaTek -- chip manufacturers with a direct interest in optical I/O standards for their own future silicon.
- Wiwynn and Alchip -- server and ASIC manufacturers who would build the physical systems around Ayar Labs' technology.
The new capital funds Ayar Labs' shift toward high-volume manufacturing -- validation, product development, and scaling the manufacturing ecosystem -- plus a new design center opening in Bengaluru, India. That transition is the hard part for any chip startup: designing a working part in a lab is one thing, getting foundries and packaging partners to produce it at the volumes hyperscalers need is another, and it's where well-funded optical-interconnect rivals like Ranovus and Ayar's own earlier-stage competitors have historically stalled.
The round adds to a week where physical AI-infrastructure financing dominated venture headlines -- Pulse also covered Fluidstack's Pentagon loan talks and Oracle's $664 billion cloud backlog in this issue alone. Optical interconnects sit further down the stack than either of those stories, but the through-line is the same: every layer of AI infrastructure, from financing to fiber, is getting re-priced upward at once.
