B300 (HGX Blackwell Ultra) specs
NVIDIA · Shipping now · first available Sep 2025 (estimated). Dense compute, memory, bandwidth and power from the vendor's documents; each precision stays separate.
All accelerators · Compare with GB300 NVL72 (per GPU)
Specifications
| Dense BF16 · FP32 accumulate | 2,250 TFLOPS |
|---|---|
| Dense FP8 · FP32 accumulate | 4,500 TFLOPS |
| Dense FP4 · FP32 accumulate | 13,500 TFLOPS |
| Memory | 270 GB · HBM |
| Memory bandwidth | 7.7 TB/s |
| Maximum chip power | 1,100 W |
| Interconnect | NVLink 5 · 1.8 TB/s |
| Process node | TSMC 4NP |
| First availability | Sep 2025est — estimated / secondary-source figure |
| Availability | Shipping |
HGX variant: dense FP4 is 108 PF per eight-GPU system divided by eight. Current datasheet specifies 270 GB and 1,100 W; GB300 NVL72 has different per-GPU specifications.
Dense tensor rates with FP32 accumulation, never sparse or FP16-accumulate headline figures. Memory is in decimal GB. The est marker flags a value its vendor-document review did not confirm.
Rent it
$7.40/GPU-hrCheapest on-demand · Hyperstack · observed
6 providers: AWS, Hyperstack, Nebius, Oracle OCI, Together AI, Verda. Quotes checked through .
B300 rental prices from every provider →Value ratios
| $ per dense BF16 PFLOP-hour | $3.29 |
|---|---|
| Dense BF16 TFLOPS per watt | 2.05 |
| Bandwidth bytes per 1,000 BF16 FLOP | 3.42 |
| $ per dense FP8 PFLOP-hour | $1.64 |
| Dense FP8 TFLOPS per watt | 4.09 |
| Bandwidth bytes per 1,000 FP8 FLOP | 1.71 |
| $ per GB of memory per hour | $0.027 |
Prices divide the cheapest on-demand list price; theoretical ratios do not measure application speed.
Sources
Specifications reviewed . Current HGX column specifies 270 GB, 7.7 TB/s and 1,100 W; older reference tables list 288 GB. Dense FP4 is 108 PF/system ÷ 8 (the per-GPU figure is rounded to 14 PF in the PDF).
Rental price: Hyperstack published pricing ↗ (NVIDIA B300 288GB; maximum per GPU: 28 pCPUs and 240 GB RAM; actual allocation not specified; published per GPU-hour).