Hardware / GPU servers / RTX Pro 6000 Blackwell
Spec sheet · Blackwell · 2025
NVIDIA RTX Pro 6000 Blackwell: Workstation GPU
Blackwell in a PCIe card: 96 GB of GDDR7 with ECC, for inference, fine-tuning and visual computing in standard servers and workstations.
Full specifications
Compute
- GPU
- NVIDIA RTX Pro 6000 Blackwell
- FP4
- ~2 PFLOPS dense
- FP8
- ~1 PFLOPS dense
Memory
- GPU memory
- 96 GB GDDR7
- Memory bandwidth
- 1.79 TB/s
Interconnect
- NVLink
- None: cards talk over PCIe
Power and cooling
- TDP
- 600 W (Workstation, Server) · 300 W (Max-Q)
Platform
- Form factor
- Dual-slot PCIe 5.0 x16
- Display
- 4× DisplayPort 2.1
- Architecture
- Blackwell
- First shipped
- 2025
Power, cooling and rack density
| Cooling | Per node | Per rack | Rack power |
|---|---|---|---|
| Air | 6.5 kW | 4 nodes | 26 kW |
A 32-node cluster (256 GPUs) needs about 0.3 MW of facility power: a colocation cage or suite. Size your own cluster in the site planner →
What one server can serve
Largest open models that fit on eight GPUs at FP8 with an 8K context, with estimated output throughput.
- gpt-oss-120b11K tok/s
- Qwen3.6 27B23K tok/s
- gpt-oss-20b30K tok/s
Cost per million tokens in the inference calculator →
How it compares
| System | Architecture | Memory / GPU | Bandwidth | FP8 dense | Max power | Cooling |
|---|---|---|---|---|---|---|
| HGX B300 | Blackwell Ultra | 288 GB | 8 TB/s | 4.5 PF | 15 kW | Liquid / Air |
| HGX B200 | Blackwell | 180 GB | 8 TB/s | 4.5 PF | 14.3 kW | Air / Liquid |
| HGX H200 | Hopper | 141 GB | 4.8 TB/s | 2.0 PF | 10.2 kW | Air / Liquid |
| HGX H100 | Hopper | 80 GB | 3.35 TB/s | 2.0 PF | 10.2 kW | Air / Liquid |
| HGX A100 80GB | Ampere | 80 GB | 2 TB/s | — | 6.5 kW | Air |
| RTX Pro 6000 Blackwell | Blackwell | 96 GB | 1.8 TB/s | 1.0 PF | 6.5 kW | Air |
Alternatives to consider
What to verify before you buy
- Verify ECC support is enabled: critical for inference workloads
- Confirm PCIe 5.0 x16 slots and power for 600 W cards (300 W for Max-Q)
- Edition: Server Edition for passively cooled servers; Workstation or Max-Q for desktops
Export controls
RTX Pro 6000 Blackwell may be subject to export controls depending on destination. Verify end-user requirements for international orders.
Frequently asked questions
How much power does an RTX Pro 6000 Blackwell draw?
Up to 6.5 kW per 8-card PCIe server at maximum rated power. Plan site power for that peak: training runs close to it. Inference typically averages 55–65% of it.
Is the RTX Pro 6000 Blackwell air-cooled or liquid-cooled?
The RTX Pro 6000 Blackwell is air-cooled.
How much GPU memory does an RTX Pro 6000 Blackwell have?
96 GB per GPU, 768 GB across eight, at 1.8 TB/s per GPU.
What models can one RTX Pro 6000 Blackwell node serve?
At FP8 with an 8K context, the largest open models that fit on eight GPUs include gpt-oss-120b, Qwen3.6 27B, gpt-oss-20b. The inference calculator shows throughput and cost for each.
What kind of site does an RTX Pro 6000 Blackwell cluster need?
A 32-node cluster (256 GPUs) needs roughly 0.3 MW of facility power with air cooling: a colocation cage or suite.
Sources: NVIDIA RTX PRO 6000 Blackwell Server Edition. Power is the maximum rated system draw; throughput and site figures are planning estimates from the same models as the Rack Rate tools.
Building with RTX Pro 6000 Blackwell?
Send us the cluster spec. We source the full bill of materials, and can finance it.