Supermicro has launched a new line of servers specifically for hyperscalers.

According to a company statement, the new MegaDC servers have a reduced component count and optimized power distribution and backplane designs.

Supermicro MegaDC 2U GPU Server (1).jpg
Supermicro's MegaDC 2U GPU Server – Supermicro

Give us the deets

MegaDC supports open standards such as OpenBMC for management, advanced I/O modules, and common redundant power supplies.

"As we continue to rapidly expand our production capacity, Supermicro is now well-positioned to service hyperscale data centers," said Charles Liang, president, and CEO of Supermicro. "MegaDC servers are optimized to reduce deployment times and deliver optimal performance per Watt and performance per dollar.

“We understand that large data centers often face long lead times for upside demand as well as occasional downside challenges, and Supermicro can help alleviate these demand fluctuation concerns by maintaining healthy inventory levels for our new MegaDC servers."

The introduction of MegaDC includes five new X11 motherboard systems comprised of two 1U systems and three 2U chasses available for large-scale cloud deployments.

DCD understands all of the MegaDC systems support two 2nd Gen Intel Xeon Scalable processors, 16 memory slots, and an AIOM slot, dual 25G Ethernet ports, and OpenBMC.

The 2U chassis has also been pitched as a machine learning platform that can accommodate multiple double-wide or single-wide GPUs configurations.

For example, two full-length, double-wide (x16) GPU cards for general-purpose GPU applications, five low-profile, single-wide GPU cards for AI applications, or two full-length, double-wide (x16) and two low-profile, single-wide GPU cards.