EME-MOTIONAL OPTICSCONNECTIVITY & PHOTONICS Request a Quote

Has the GB200 abandoned optical modules

The GB200 has not abandoned optical modules; it currently uses OSFP optical modules for NVLink expansion, though future designs are moving toward Co-Packaged Optics (CPO).

Current Use of Optical Modules

The NVIDIA GB200 NVL72 systems employ OSFP optical modules as expansion interfaces. Each OSFP module supports 16 differential signal lines, enabling 4 NVLink ports per module. A single NVLink switch in the GB200 contains 32 OSFP connectors, cumulatively supporting 128 NVLink 4 ports . These optical modules are integral to the system's high-bandwidth interconnect, allowing each GPU pair to achieve the 1.8 Tbps aggregate bandwidth required by NVLink 5.0 . The external interconnect of the NVL72 rack, however, uses direct copper backplane connections for intra-rack communication, while optical modules are primarily used for expansion and scaling .

Transition Toward Co-Packaged Optics (CPO)

While OSFP modules remain standard in current GB200 deployments, NVIDIA is exploring CPO technology to address thermal and power challenges in large-scale deployments. CPO integrates optical engines directly with the switch ASIC, reducing power consumption and improving signal density. Mass production of CPO-based switches is expected to support up to 115.2 Tbps signal transmission, with trial production already underway . This indicates a gradual shift toward CPO for future GB200 systems, but optical modules have not been fully abandoned.

Rationale

The continued use of OSFP optical modules is driven by:

  • Bandwidth requirements: 800G modules cannot meet the 1.8 Tbps per GPU pair needed for Blackwell GPUs, necessitating 1.6T optical transceivers .
  • System compatibility: Current GB200 racks ship with 1.6T optical ports, and adapters for older 800G modules introduce unacceptable latency .
  • Scalability: Optical modules allow flexible expansion of NVLink ports across multiple racks without redesigning the copper backplane .

Conclusion

In summary, GB200 systems continue to use OSFP optical modules for NVLink expansion, particularly for high-bandwidth interconnects. The industry is moving toward CPO solutions to improve efficiency and thermal performance, but this represents a future evolution rather than a complete abandonment of optical modules .

Has the GB200 abandoned optical modules  - E-Motional Optics & Connectivity

GB200 NVL72 | NVIDIA

Discover the powerful GB200 NVL72 GPU, engineered for AI workloads and next-gen data centers.

Unwrapping the NVIDIA B200 and GB200 AI GPU Announcements

NVIDIA also announced the Grace-Blackwell GB200 Superchip. This is a module that has two B200 GPUs wired to

GB200 Hardware Architecture

With CX-7 on the GB200 NVL72, each GPU has 400G of bandwidth and is connected to one OSFP cage with the

Huawei AI CloudMatrix 384 – China''s Answer to Nvidia GB200 NVL72

Huawei AI CloudMatrix 384 – China''s Answer to Nvidia GB200 NVL72 China Abundance of Power, 100% Optics, 0%

NVIDIA GB200 User Guide: Specs, Features and Use Cases

Learn about the NVIDIA Blackwell GB200. Enterprise-scale AI with superior performance, advanced networking and

Q&A with Nvidia''s Chief of DGX Systems on the DGX-GB200

Pictures of Nvidia''s new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on

NVIDIA GB200 Interconnect Architecture Analysis: NVLink, InfiniBand

In this process, high-speed interconnect devices such as optical modules will continue to play a crucial role in providing

GB200 order update: Microsoft''s demand exceeds total orders

Conclusions: Microsoft''s demand for GB200 surpasses the total orders from other CSPs, with 4Q24 orders

Nvidia Orders Surge: InnoLight and Eoptolink Dominate 60% of 800G

Following the 800G optical modules, as AI server clusters demand higher interconnection speeds, Nvidia has chosen

NVIDIA GB200: Interconnect Architecture and Evolution

Each GB200 subsystem has 2 * 18 = 36 NVLink5 Ports. The external interconnect of the

Understanding Your Grace-Blackwell Systems — NVIDIA GB200 NVL

Understanding Your Grace-Blackwell Systems # GB200 NVL is a rack scale solution for GPUs connected using

Why 1.6T Optical Transceivers Overtake 800G in 2026 AI Clusters?

1.6T optical transceivers are overtaking 800G in 2026 AI clusters because NVIDIA Blackwell GB200 systems require doubled

The NVIDIA Grace Blackwell Superchip — NVIDIA GB200 NVL Multi

The GB200 Grace Blackwell Superchip is a key component of the NVIDIA GB200 NVL72, connecting two high

800G Optical Transceivers Clear the Path for NVIDIA

The fabric requirements for GB200 deployments push optical transceivers on latency, density, thermal

Nvidia GB200 production ramps up after suppliers tackle AI server

Nvidia suppliers building its Blackwell AI server racks have reportedly solved a series of technical hurdles, allowing

NVIDIA GB200: Inside the Blackwell Superchip Powering AI at Scale

When NVIDIA unveiled the GB200 Grace Blackwell Superchip, I realized we were witnessing the next evolution of AI compute. As

Hardware — NVIDIA DGX GB Rack Scale Systems User Guide

Multiple power shelves are used in the rack to supply redundant power. The rack power consumption is approximately

Still Have a Technical Question?

Our photonic engineering team can help you select the right connector or splitter for your network.

Ask Our Team