NVIDIA RTX PRO 6000 96GB GDDR7 AI Accelerator Card 900-5G144-2200-000
- RTX PRO 6000
- PCIe 5.0 x16
We take cryptocurrency — so the coins you hold can buy real hardware.Pay in crypto, spend it on real hardware. Free insured shipping worldwide over $399. Plain, unbranded boxes, sent from Germany.Free insured shipping over $399. Your rate is locked for 30 minutes once checkout opens.Rate locked 30 minutes. Every item sealed, serial-checked and covered by a 24-month warranty.Sealed, 24-month warranty.
64 GB HBM2 AI accelerator with 6,656 stream processors and 181 TFLOPs peak FP16 performance
The Supermicro GPU-XGMIMI210-4P is a new AMD Radeon Instinct MI210 accelerator built for enterprise HPC and AI workloads. It delivers 64 GB of HBM2 memory and 6,656 stream processors organised in 104 compute units. Peak half-precision throughput reaches 181 TFLOPs at a 1,700 MHz boost clock.
This card targets researchers and enterprises running large-model training or inference on a single server. The 300 W dual-slot design fits standard PCIe slots. Buyers who need multi-accelerator scaling or workloads that exceed 64 GB memory should evaluate alternative configurations.
The 64 GB HBM2 frame buffer is the single specification that determines whether this accelerator matches a workload. Models and datasets that fit within this capacity benefit from high-bandwidth memory; larger problems will require cards with more memory or distributed solutions.
Measured against everything we stock in GPUs for AI & Data Centers, at today’s prices. Recalculated every night.
New to this? Buying server hardware with crypto — what to check, what it costs to pay, and what is never asked for.
| Size & build | |
|---|---|
| Shipping weight | 4.54 kg |
| Package size | 254 × 254 × 254 mm |
The card draws up to 300 W under full load, so sustained HPC or training jobs will generate significant heat that requires adequate server airflow and cooling capacity.
A PCIe x16 slot with sufficient physical clearance, a power supply able to deliver 300 W to the card, and chassis airflow rated for high-density compute modules are required.
The accelerator communicates over standard PCIe, so it functions with existing CPUs and system memory as long as the motherboard provides a compatible x16 link and the platform BIOS supports add-in compute devices.
More in GPUs for AI & Data Centers
What you pay in, and what you read prices in.
Sent from your wallet at checkout
Display only — the catalogue is priced in USD
No currency matches that.
New here?
Already have one?
There is no automatic reset: this shop sends no mail, so there is no one-time link to ask for. Leave the address on the account and it gets picked up by hand.
Request noted
Nothing is on its way to your inbox — this shop sends no mail at all. Someone picks these up by hand.
Remembered it?