Multi-node servers with shared power and cooling for high-density compute and HCI clusters.
Supermicro Twin systems place several independent server nodes in one 2U chassis that share redundant Titanium-level power supplies and cooling. This raises compute density and lowers cost and power per node.
BigTwin provides four dual-processor nodes in 2U for performance-focused clusters, while GrandTwin provides four single-processor nodes with front or rear I/O for scale-out cloud, HCI and HPC.
Independent hot-pluggable nodes in a shared chassis.
Redundant high-efficiency power supplies serve all nodes.
Intel Xeon or AMD EPYC 9005/9004 with up to 192 cores per node.
DDR5 memory on current GrandTwin models.
GrandTwin options for different data center layouts.
| Nodes | 2U, 4 single-socket nodes |
|---|---|
| CPUs | AMD EPYC 9005 / 9004 (up to 192 cores) or Intel Xeon, per model |
| Memory | Up to 4 TB DDR5 per node |
| Storage | 4 to 6 hot-swap NVMe / SATA bays per node |
| Power | 2200 W redundant Titanium (model dependent) |
| Nodes | 2U, 4 dual-socket nodes |
|---|---|
| Storage | Up to 24× 2.5" or 12× 3.5" NVMe / SAS / SATA per chassis |
| Power | Redundant Titanium-level power supplies |
| AS -2116GT-HNTF | GrandTwin 2U 4-node, AMD EPYC 9005 / 9004 |
|---|---|
| AS -2115GT-HNTR | GrandTwin 2U 4-node, rear I/O, 6 bays per node |
| SYS-211GT-HNC8F | GrandTwin 2U 4-node, Intel Xeon Scalable |
Flagship 1U/2U dual-processor rackmount servers for virtualization, databases and enterprise apps.
Cost-optimised single- and dual-socket servers for cloud and hosting providers.
Blade architectures with integrated switching for maximum density and energy efficiency.
Tell us your workload and timeline for the Supermicro BigTwin & GrandTwin. We reply within one business day with a configuration, pricing and lead time.