DELL
DELL NVIDIA A2 Tensor Core 16GB GDDR6 60W HHHL SW LP PCIe GPU
NVIDIA A2 Tensor Core GPU with 16 GB GDDR6 memory and 60 W power draw for low-profile server inference
- TypeRack
- 16GB GDDR6 memory powers AI inference workloads
- 60W low-profile design fits dense server racks
- Tensor Core architecture accelerates deep-learning tasks
- Dell 490-BHMD part ensures compatibility with supported systems
Dell NVIDIA A2 Tensor Core GPU
The Dell NVIDIA A2 Tensor Core GPU provides 16GB of GDDR6 memory for AI inference workloads in server environments. Its 60W power envelope suits dense rack deployments where thermal headroom is limited. The half-height half-length low-profile form factor fits 1U and 2U chassis without blocking adjacent slots. This card targets inferencing rather than training pipelines.
PCIe low-profile integration needs
The GPU installs into a standard PCIe slot and draws up to 60W from the bus, so the host board must supply that power without auxiliary connectors. Low-profile brackets are required for the target chassis height. The single-slot design leaves neighbouring PCIe lanes free for networking or storage controllers. Buyers who need active cooling fans or higher double-precision throughput should consider other accelerator classes.
Highlights
- 16GB GDDR6 memory powers AI inference workloads
- 60W low-profile design fits dense server racks
- Tensor Core architecture accelerates deep-learning tasks
- Dell 490-BHMD part ensures compatibility with supported systems
Specifications
| Type | Rack |
|---|
Questions about this item
Does the 60W board-power limit throttle sustained inference?
The 60W ceiling is the design limit; the low-profile heatsink and blower keep clocks steady at that draw.
How loud and hot does the card get under all-day load?
The single blower exhausts heat out the rear bracket; at 60W it runs warm but within server thermal specs.
Also in this aisle
People compared these
Same shelf, same checkout — eight coins and a 30-minute rate lock.



