FS Delivers 1.6T OSFP IHS/Closed Finned Top Modules for the Next AI/HPC Network

FS has unveiled its new 1.6T OSFP IHS (closed finned top) optical transceivers, including the OSFP-DR8-1.6T and OSFP-2FR4-1.6T models for high‑density AI and HPC InfiniBand XDR networks. Designed to deliver full 1.6T throughput into next‑generation GB300/B300 architectures, these modules represent a transformative solution for organizations seeking to harness the full potential of AI without compromising on efficiency or compatibility. By breaking through traditional performance bottlenecks, FS Inc. empowers global enterprises to equip their infrastructures with the scalable power needed to solve the most complex computational challenges.

As AI infrastructures evolve from rack-scale to row-scale deployments, link distances, east‑west traffic, and power constraints are all increasing. Traditional copper‑based interconnects are reaching their physical limits in terms of reach and signal integrity. At the same time, next‑generation compute platforms such as NVIDIA GB300/B300 are driving a hard requirement for 1.6T optical interconnects. Against this backdrop, pluggable 1.6T optical interconnect solutions have become a key building block for AI/HPC network upgrades, enabling a smoother transition beyond 800G while meeting the performance and power targets of large‑scale AI clusters.

FS 1.6T OSFP Modules: High Bandwidth, Low Power, Proven for InfiniBand AI Fabrics

The FS 1.6T OSFP IHS transceiver series is engineered to balance performance and energy efficiency for next‑generation AI infrastructures.

Features

• Rigorous Testing: FS 1.6T optical modules are 100% verified on NVIDIA Quantum‑X800 switches in InfiniBand XDR environments, delivering a plug-and-play, out of the box installation experience with stable performance. Each module is qualified for low BER, high signal integrity, and strong durability, helping enterprises minimize link failures and reduce long-term operation and maintenance costs.
• High Bandwidth: Utilizing 8x 200G PAM4 modulation, the FS 1.6T OSFP IHS modules provide much higher bandwidth density than mainstream 800G transceivers, enabling enterprises to build larger AI and HPC clusters within the same rack space and accelerate large‑scale AI training.
• Low Power Consumption: Powered by the latest Broadcom 3nm DSP, the series significantly reduces power consumption (as low as 26W). The OSFP-DR8-1.6T further integrates Silicon Photonics (SiPh) technology, minimizing thermal loads and cooling pressure. This allows data centers to scale their AI workloads while maintaining lower energy expenditures and operational costs.
• Flexible Deployment Scenarios: Featuring an integrated heat sink (IHS) with cooling fins, these transceivers are compatible with both air-cooled and liquid-cooled environments, enabling greater flexibility in modern data center architectures.
  • OSFP‑DR8‑1.6T (up to 500m): This module is primarily used for interconnection between 1.6T NVIDIA Quantum-X800 and NVIDIA B300/GB300 systems, achieving high-speed interconnection in AI data centers and cloud computing areas. It helps enterprises simplify cabling and reduce long‑term operation and maintenance costs.
  • OSFP‑2FR4‑1.6T (up to 2km): With CWDM4 wavelength‑division multiplexing, the OSFP‑2FR4‑1.6T supports the longer connection between NVIDIA Quantum‑X800 InfiniBand switches, enabling 1.6T high‑bandwidth links for large AI clusters and trillion‑parameter AI models.

“The evolution from 800G to 1.6T is not only about higher speed but also about reshaping the optical interconnect architecture for AI,” said Kyrie Zhang, Senior Product Manager at FS Inc. “FS is committed to delivering proven and ready‑to‑deploy 1.6T solutions, and our 1.6T OSFP IHS transceiver series gives global professional organizations the reliability and efficiency needed to overcome performance bottlenecks and scale massive clusters.”

In the era of large-scale HPC and AI, the FS 1.6T InfiniBand OSFP IHS transceivers combine 1.6T bandwidth, low power consumption, and enhanced thermal performance to support today’s rack-scale and row-scale AI clusters. With InfiniBand XDR compliance, verified interoperability on NVIDIA Quantum X800 platforms, and an expanding 1.6T portfolio, the FS 1.6T OSFP series is ready to power the next wave of high-density AI and HPC networks.

More information about the OSFP-DR8-1.6T models from FS

More information about the OSFP-2FR4-1.6T models from FS