Scale AI Infrastructure with 800G and 1.6T
Production-ready 800G and 1.6T optical and high-speed copper connectivity, designed for real AI deployments - from today’s clusters to next-generation fabrics.
Talk to an ATOP Specialist
Production-Ready Connectivity. Not Concept Hardware.
ATOP’s 800G and 1.6T portfolio is engineered for real-world deployment, not lab-only demonstrations.
Our solutions span both optical and copper interconnects, designed to meet the performance, power, and reliability demands of dense GPU and accelerator clusters.
Every product is built with deployment conditions in mind - from interoperability and signal integrity to manufacturability at scale.
Every product is built with deployment conditions in mind - from interoperability and signal integrity to manufacturability at scale.
1.6T in Practice
1.6T OSFP DR8 LRO
Designed for ultra-low latency AI fabrics where predictable performance and power efficiency are critical.
1.6T OSFP AEC (3m)
High-speed copper interconnects optimized for short-reach, power-efficient AI infrastructure.
Deployable, production-ready 1.6T
Why 1.6T LRO Matters for AI Infrastructure
ATOP’s 1.6T OSFP DR8 SiPh LRO is engineered for modern AI and hyperscale environments where bandwidth, power efficiency, and latency must scale together.
• Hot-pluggable OSFP form factor
• Aggregate bandwidth of 1.6 Tbps
• Eight lanes of 212.5 Gb/s PAM4
• 3nm DSP architecture
• Silicon photonics transmitters at 1310 nm
By removing receive-side DSP, the LRO architecture reduces processing overhead, minimizes latency, and eliminates GPU wait-states in tightly coupled AI fabrics.
• Aggregate bandwidth of 1.6 Tbps
• Eight lanes of 212.5 Gb/s PAM4
• 3nm DSP architecture
• Silicon photonics transmitters at 1310 nm
By removing receive-side DSP, the LRO architecture reduces processing overhead, minimizes latency, and eliminates GPU wait-states in tightly coupled AI fabrics.
Designed for AI at Scale
Ultra-Low Latency by Design
Lower Power per Terabit
Predictable Performance at Scale
What This Enables at Scale
As AI clusters grow, latency consistency and power efficiency become just as critical as raw bandwidth.
ATOP’s LRO architecture supports:
• Higher port density
• Stable performance under sustained workloads
• Faster time-to-compute
• Improved efficiency across training and inference environments
This is connectivity designed for scale, not peak benchmarks.
• Higher port density
• Stable performance under sustained workloads
• Faster time-to-compute
• Improved efficiency across training and inference environments
This is connectivity designed for scale, not peak benchmarks.
AI-Ready Connectivity, from 800G to 1.6T
ATOP delivers end-to-end 800G and 1.6T connectivity, supporting AI infrastructure across the full network lifecycle.
Customers can deploy 800G today while planning a seamless transition to 1.6T, without forced upgrades or supply uncertainty.
Our portfolio supports:
• Optical transceivers
• High-speed copper interconnects
• AI training, inference, hyperscale, and HPC environments
Customers can deploy 800G today while planning a seamless transition to 1.6T, without forced upgrades or supply uncertainty.
Our portfolio supports:
• Optical transceivers
• High-speed copper interconnects
• AI training, inference, hyperscale, and HPC environments
Powering More Sustainable AI Infrastructure
Power efficiency is now a defining requirement for AI at scale.
ATOP’s 800G and 1.6T copper solutions help reduce:
• Power consumption
• Infrastructure complexity
• Short-reach inefficiencies
Optimized interconnects contribute to lower energy usage and support customers balancing performance with long-term sustainability goals.
• Power consumption
• Infrastructure complexity
• Short-reach inefficiencies
Optimized interconnects contribute to lower energy usage and support customers balancing performance with long-term sustainability goals.
Why It Matters at Scale
Lower Power per Link
Optimized for Short-Reach AI Fabrics
Reduced Infrastructure Footprint
Scalable, Traceable Supply, Built for Growth
ATOP designs and manufactures with scalability, traceability, and compliance at the core.
• Certified and validated products
• Automated manufacturing processes
• Consistent quality at volume
• Actionable sustainability and ESG data
With global production capacity, ATOP supports large-scale 800G and 1.6T deployments worldwide.
• Automated manufacturing processes
• Consistent quality at volume
• Actionable sustainability and ESG data
With global production capacity, ATOP supports large-scale 800G and 1.6T deployments worldwide.