a

AI Scale-up Beyond the Rack

Ayar Labs’ co-packaged optics (CPO) for AI scale-up. Built and backed by semiconductor ecosystem leaders.

Explore Our CPO Solution

Purpose-Built to Scale AI

Ayar Labs’ co-packaged optics (CPO) solution unlocks AI performance and profitability by enabling thousands of GPUs to operate as a single unified system with the bandwidth and latency needed for hyperscale AI infrastructure.

Proven and Built for High Volume Manufacturing

3

generations of optical I/O silicon

1e-12

Bit Error Rate (BER)

800°C/s

validated thermal slew rate

Ayar Labs’ TeraPHY™ optical engine is built using the standard form factor, manufacturing, and packaging flows already utilized by major XPU and switch vendors, making it easy to integrate into existing customer designs.

Backed by leading industry players, validated across multiple generations,and supported by extensive verification and reliability testing, Ayar Labs’ TeraPHY optical engine is purpose-built and production-ready for hyperscale infrastructure demands.

Integrated into the Package

>8 Tbps

per optical engine

<25 ns

end-to-end optical I/O latency

>5 dB

optical margin across links

Purpose-built for flexible integration. TeraPHY™ optical engines integrate directly into existing package architectures, replacing copper’s limitations with high-bandwidth, low-latency optical connectivity and protocol-agnostic flexibility.

Inside the Rack

16+

SuperNova™ ELSFPs per tray

32

trays per rack

1000s

of GPU connections

Maximize every millimeter and watt of your AI infrastructure. Our optical engines provide the bandwidth density and power efficiency needed to pack more AI performance into every rack while eliminating bottlenecks.

Our SuperNova™ remote light source, with its compact, low-power design in the industry-standard ELSFP form factor pairs seamlessly with TeraPHY to deliver a complete, manufacturing-ready solution for optically-connected rack-scale architecture.

Beyond the Rack

Enable hyperscale AI. Ayar Labs’ CPO solutions connect thousands of GPUs across multiple racks into a single unified cluster via optical fabric, delivering the connectivity hyperscalers need to unleash next-generation AI performance.

See our Scale-Up Solution

Ready to Scale Beyond the Rack

Whether you’re designing next-generation silicon, scaling AI systems, or exploring CPO, our team is ready to work with you.

Contact Us