NVIDIA GB300 GPU: Specs, VRAM, Price & Benchmarks (2026)
The GB300 is a real GPU (launched 2025), but no provider on Aquanode is listing it for rental right now, so there is no live hourly price to quote. Availability changes as providers add and retire hardware; the specs, VRAM and price context below still apply.
How much VRAM does the GB300 have?
The GB300 has Up to 1TB unified per Superchip (2x 288GB HBM3e Blackwell Ultra GPUs + Grace CPU LPDDR5X), with 8 TB/s per GPU (HBM3e) of peak memory bandwidth.
GB300 specs
| Architecture | launched 2025 |
| VRAM | Up to 1TB unified per Superchip (2x 288GB HBM3e Blackwell Ultra GPUs + Grace CPU LPDDR5X) |
| Memory bandwidth | 8 TB/s per GPU (HBM3e) |
| Interconnect | NVLink-C2C, 900 GB/s CPU-GPU |
| TDP | Up to 1,400W per Blackwell Ultra GPU (liquid-cooled Superchip module) |
| Form factor | Superchip module (rack-scale, liquid-cooled; deployed in the GB300 NVL72 rack) |
Specs sourced from the vendor's public product page. See the source.
GPU Glossary: What is VRAM?, HBM, Tensor Cores, CUDA Cores, TFLOPS, NVLink vs PCIe
Good for
The Blackwell Ultra follow-on to GB200: 50% more HBM3e per GPU (288GB vs 192GB) in the same rack-scale NVL72 form, aimed at trillion-parameter-class training and the highest-throughput reasoning-model inference. Like GB200, this is a complete rack system, not a per-GPU rental product; Aquanode's live Blackwell pricing above (B200) is the closest per-GPU-hour equivalent available today.
Not good for
Same as GB200: not available as a single card or a per-GPU-hour rental anywhere, including on Aquanode. A full NVL72 rack is the unit of deployment.
Related guides
Other models in the same generation. The full list is in the GPU index.
Get notified when the price drops
GPU supply moves hourly. Tell us what you're waiting for and we'll email you when a matching offer appears across any provider we track.
