NVIDIA MFS1S50-H010E 200G QSFP56 to 2x100G QSFP56 AOC, 10 m | Low Latency MMF Splitter Cable

The NVIDIA MFS1S50 series is a family of QSFP56 VCSEL-based active optical splitter cables designed as cost-effective 200Gb/s to 2x100Gb/s interconnects for 200Gb/s InfiniBand HDR and 200GbE systems. The MFS1S50-H010E variant provides a 10-meter reach and connects a single 200Gb/s QSFP56 port on one side to two separate 100Gb/s QSFP56 connectors on the other — for example, a switch uplink to two servers.

The cable routes data signals from each of the four multimode fiber (MMF) pairs on the single QSFP56 end to the dual pair of each QSFP56 multiport end. Each QSFP56 end includes an EEPROM that provides product and status monitoring information readable by the host system. The cable is SFF-8665 compliant for the QSFP56 pluggable solution and supports the SFF-8636 I²C management interface.

NVIDIA cable solutions are designed for power-efficient data center interconnects, enabling higher port bandwidth, density, and configurability at reduced power requirements. Rigorous production testing supports out-of-the-box installation, performance, and durability.

• Supports IBTA IB HDR and 200GbE
• 200Gb/s HDR to 2x100Gb/s HDR100 data rate
• 200GbE to 2x100GbE data rate
• 4x 50Gb/s PAM4 modulation
• Programmable Rx output amplitude and pre-emphasis
• SFF-8665 compliant QSFP56 port
• Single 3.3V power supply
• 4.5 W typical power consumption at the 200G end
• Up to 30 m length available in the series; this variant is 10 m
• Hot pluggable
• RoHS compliant
• SFF-8636 compliant I²C management interface

The cable uses VCSEL (Vertical Cavity Surface-Emitting Laser) technology over multimode fiber. The 200Gb/s end transmits four lanes at 50Gb/s PAM4, and the splitter architecture maps these lanes to two 100Gb/s QSFP56 ends, each carrying two lanes. This allows a single high-density 200Gb/s switch port to serve two 100Gb/s endpoints without requiring separate 200G optics on each server side.

The AOC supports programmable Tx input equalization, programmable Rx output amplitude, and programmable Rx output pre-emphasis for physical layer link optimization. Tx/Rx CDR control is provided; CDR is enabled by default for 100GbE operation and can be disabled for 40G operation. Digital Diagnostic Monitoring (DDM) covers Rx optical power per lane, Tx optical power per lane, Tx bias current per lane, supply voltage, and transceiver case temperature, with warning and alarm thresholds.

Rate select is supported via registers 0x57–0x58. Two bits are assigned per receiver lane in byte 0x57 and two bits per transmitter lane in byte 0x58 to specify up to four bitrates per SFF-8636 Rev 2.9.2 Table 6-5 XN_RATE_SELECT ENCODINGS. All four lanes must use the same rate select value.

• 200Gb/s InfiniBand HDR switch port fan-out to two 100Gb/s HDR100 server ports
• 200GbE switch uplink split to two 100GbE server or storage ports
• Data center leaf-spine architectures requiring 200G-to-2x100G breakout
• High-performance computing (HPC) clusters using HDR fabrics
• Rack-to-rack and row-level optical interconnects within the 10 m reach

The MFS1S50-H010E is designed for QSFP56 ports supporting 200Gb/s InfiniBand HDR and 200GbE. The 200Gb/s end connects to a single QSFP56 port; the 100Gb/s ends connect to two QSFP56 ports operating at 100Gb/s. It is SFF-8665 and SFF-8636 compliant. For non-NVIDIA hosts, compatibility and breakout support should be confirmed before ordering.

The MFS1S50 series is available in multiple lengths from 3 m to 30 m. The table below lists the ordering part numbers from the product specification to help select the correct reach.

Note: The ordering table in the product specification lists 10 m as MFS1S50-H010V. This page covers the MFS1S50-H010E designation; please confirm the exact part number suffix with sales before ordering.

• Enables 200G-to-2x100G breakout from a single QSFP56 switch port
• Low latency active optical design for HDR and 200GbE fabrics
• Supports HDR to 2xHDR100 and 200GbE to 2x100GbE configurations
• SFF-8665 and SFF-8636 compliant for management and interoperability
• Programmable Tx equalization and Rx amplitude/pre-emphasis for link tuning
• DDM monitoring of optical power, bias current, voltage, and temperature
• Hot pluggable and RoHS compliant
• Up to 30 m reach available in the series for flexible rack layouts

Hong Kong Starsurge Group Co., Limited supplies NVIDIA / Mellanox LinkX cables and related network hardware to customers worldwide. Our sales and technical team can assist with part number selection, length decisions, compatibility verification, and quotation requests. We support industries including government, healthcare, manufacturing, education, finance, and enterprise.

Please confirm current lead time, MOQ, shipping terms, and the exact part number suffix with our sales team before ordering.

• Observe the 30 mm minimum bend radius and 3 ± 0.2 mm cable diameter during routing.
• Keep the optical port dust cap in place when no fiber cable is connected.
• Clean the fiber cable connector before insertion to prevent contamination.
• Use standard ESD handling precautions, including grounded wrist/shoe straps and ESD flooring, when inserting or extracting the cable.
• Do not exceed the absolute maximum ratings for supply voltage, data input voltage, or control input voltage.
• Operate within the 0°C to 70°C case temperature and 5% to 85% relative humidity range.
• Confirm the exact part number suffix and host breakout support before ordering.

Hong Kong Starsurge Group Co., Limited is a technology-driven provider of network hardware, IT services, and system integration solutions. Founded in 2008, the company serves customers worldwide with products including network switches, NICs, wireless access points, controllers, cables, and related networking equipment.

Backed by an experienced sales and technical team, Starsurge supports industries such as government, healthcare, manufacturing, education, finance, and enterprise. The company also offers IoT solutions, network management systems, custom software development, multilingual support, and global delivery.

With a customer-first approach, Starsurge focuses on reliable quality, responsive service, and tailored solutions that help clients build efficient, scalable, and dependable network infrastructure.

• Confirm required reach: 10 m for this variant; 3 m to 30 m options exist in the series.
• Confirm the exact part number suffix (H010E vs. H010V) with sales before ordering.
• Verify host port type is QSFP56 and supports 200G-to-2x100G breakout.
• Check whether the host requires vendor-specific coding or breakout configuration.
• Confirm operating temperature and humidity range match the deployment environment.
• Review the 30 mm minimum bend radius and cable routing plan.
• Confirm lead time, MOQ, and shipping terms with the supplier.

• MFS1S50-H003V — 3 m, 200Gb/s QSFP56 to 2x100Gb/s QSFP56 AOC
• MFS1S50-H005V — 5 m, 200Gb/s QSFP56 to 2x100Gb/s QSFP56 AOC
• MFS1S50-H015V — 15 m, 200Gb/s QSFP56 to 2x100Gb/s QSFP56 AOC
• MFS1S50-H020V — 20 m, 200Gb/s QSFP56 to 2x100Gb/s QSFP56 AOC
• MFS1S50-H030V — 30 m, 200Gb/s QSFP56 to 2x100Gb/s QSFP56 AOC

• How to plan 200G-to-2x100G breakout cabling for InfiniBand HDR and 200GbE
• Active optical cable vs. direct attach copper for 200Gb/s data center links
• Understanding SFF-8636 DDM and rate select for QSFP56 AOC cables
• Installation and cleaning best practices for MMF active optical cables
• Cable length and bend radius planning for high-density switch uplinks