Date
10/14/2025
PDF
Xi Wang, senior vice president and general manager, Connectivity Business Unit at Marvell
Marvell Technology, Inc. announced that it is expanding its industry-leading connectivity portfolio with the addition of Marvell® active copper cable (ACC) linear equalizers.
The scale and complexity of today’s AI workloads are driving exponential growth in data center bandwidth, requiring new challenges in managing thermal and power efficiency. Copper remains the preferred solution for in-rack scale-up interconnects due to its low cost and ease of deployment. However, next-generation AI systems demand thinner copper-based interconnects within server racks to improve airflow and cooling. Meanwhile, as bandwidth and cable gauge requirements continue to rise, the signal transmission performance of direct attach copper (DAC) technology is increasingly limited.
Analog ACC devices incorporate a signal equalizer, offering longer reach than traditional passive DAC cables while adding minimal latency. They are also more cost-effective and power-efficient than digital alternatives.
Leveraging Marvell industry-leading PAM4 technology and expertise in 100G/lane and 200G/lane analog devices, the new Marvell ACC linear equalizers deliver superior gain, extending the reach of ACC compared to competitive ACC solutions at the same cable gauge. They support 800G and 1.6T copper interconnects and expand the Marvell scale-up interconnect portfolio, which includes chipsets for active electrical cables (AEC) and active optical cables (AOC).
“Offering a full complement of ACC, AEC and AOC silicon technologies, Marvell is unique in the scale-up interconnect landscape, providing customers with a full range of solutions to meet their individual requirements,” said Xi Wang, senior vice president and general manager, Connectivity Business Unit at Marvell. “We are excited to work with our ecosystem of cable OEM partners and system vendors to provide end customers with high-performance, in-rack connectivity solutions to handle their most advanced AI workloads.”
For more information, go here.
