你的購物車是空的

${{ (item.variation.media ? item.variation.media.alt_translations : item.product.cover_media.alt_translations) | translateModel }}$ ${{ (item.variation.media ? item.variation.media.alt_translations : item.product.cover_media.alt_translations) | translateModel }}$

{{ childProduct.title_translations | translateModel }}

{{ getChildVariationShorthand(childProduct.child_variation) }}

{{ getSelectedItemDetail(selectedChildProduct, item).childProductName }} x {{ selectedChildProduct.quantity || 1 }}

{{ getSelectedItemDetail(selectedChildProduct, item).childVariationName }}

{{item.quantity}}x HK$0 {{ item.unit_point }} 點

刪除

${{addonItem.product.cover_media.alt_translations | translateModel}}$

{{addonItem.quantity}}x {{ mainConfig.merchantData.base_currency.alternate_symbol + "0" }}

刪除

NVIDIA MCA7K10 980-9IAO5-00X01A Active copper splitter cable, 1600Gb/s to 2x800Gb/s, OSFP to 2xOSFP, 1.5m, IHS to RHS

The NVIDIA MCA7K10 (980-9IAO5-00X01A) is a 1.5m Active Copper Splitter Cable for AI networks. It splits a 1600Gb/s OSFP port into two 800Gb/s OSFP lines, utilizing an IHS-to-RHS thermal design to link Quantum-3 switches to ConnectX-8 adapters at ultra-low latency.

全店，Free Shipping

{{shoplineProductReview.avg_score}}( {{shoplineProductReview.total}} {{'product.product_review.reviews' | translate}} )

{{amazonProductReview.avg_rating}}( ({{amazonProductReview.total_comment_count}} {{'product.product_review.reviews' | translate}}) )

一次最大商品購買數量限制為 99999

該數量不適用，請填入有效的數量。

售完

若想購買，請聯絡我們。

商品描述

送貨及付款方式

顧客評價

商品描述

NVIDIA MCA7K10 Active Copper Splitter Cable

Name: NVIDIA MCA7K10 980-9IAO5-00X01A Active copper splitter cable, 1600Gb/s to 2x800Gb/s, OSFP to 2xOSFP, 1.5m, IHS to RHS
Brand: NVIDIA
SKU: 6a460002e47171186256f638
Availability: InStock

1600Gb/s to 2x800Gb/s | OSFP to 2xOSFP | 1.5m | IHS to RHS

Product Overview

The NVIDIA MCA7K10 (MPN: 980-9IAO5-00X01A) is an ultra-high-performance Active Copper Splitter Cable (ACC) engineered for next-generation, large-scale AI clusters and high-performance computing (HPC) environments. Delivering a massive aggregate throughput of 1600Gb/s (1.6Tb/s), this premium 1.5-meter breakout cable splits a single, twin-port 1600Gb/s OSFP uplink into two independent 800Gb/s OSFP channels.

Utilizing advanced 8-channel 200G-PAM4 modulation, this active copper solution provides power-efficient, ultra-low-latency connectivity, acting as the ideal infrastructure bridge between high-density switches and next-gen network interface cards.

Key Features

1.6Tb/s High-Density Breakout: Seamlessly splits a 1600Gb/s host link into twin 800Gb/s target channels to maximize switch port density and network configurability.
Dual-Protocol Flexibility: Smart firmware automatically detects and supports both InfiniBand (XDR) and Ethernet protocols depending on the host switch configuration.
Optimized IHS-to-RHS Thermal Architecture: Designed with an Integrated Heat Sink (IHS / Closed Finned Top) on the 1600G switch side and a Riding Heat Sink (RHS / Flat Top) on the two 800G adapter ends for flawless thermal management.
Active Power Efficiency: Active signal conditioning ensures exceptional signal integrity over a 1.5-meter reach while consuming significantly less power and cost than alternative optical transceiver setups.

Technical Specifications

Parameter	Specification
Manufacturer	NVIDIA
Part Number (MPN)	980-9IAO5-00X01A
Model	MCA7K10
Cable Type	Active Copper Splitter Cable (ACC / Active DAC)
Data Rate	1600Gb/s to 2x 800Gb/s (8x 200G-PAM4)
Connector Type A	1x Twin-port OSFP224 (Host / 1600G)
Connector Type B	2x Single-port OSFP224 (Target / 800G)
Thermal Profile	Closed Finned Top (IHS) to Flat Top (RHS)
Length	1.5 Meters (5 feet)
Wire Gauge	26 AWG
Compliance	OSFP MSA, CMIS, IEEE 802.3ck, RoHS

Hardware Compatibility & Applications

Quantum to ConnectX-8 Interconnect: Explicitly tailored to link an 1600Gb/s NVIDIA Quantum-3 InfiniBand switch (finned cage) directly to two 800Gb/s ConnectX-8 PCIe network adapter cards (riding heat sink cages).

Accelerated AI Workloads: Ideal for high-density GPU clustering, LLM training pipelines, and hyperscale cloud infrastructure requiring continuous multi-terabit bandwidth.