Instinct MI355X specs
AMD · 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 Instinct MI325X
Specifications
| Dense BF16 · FP32 accumulate | 2,500 TFLOPS |
|---|---|
| Dense FP8 · FP32 accumulate | 5,000 TFLOPS |
| Dense FP4 · FP32 accumulate | 10,100 TFLOPS |
| Memory | 288 GB · HBM |
| Memory bandwidth | 8 TB/s |
| Maximum chip power | 1,400 W |
| Interconnect | Infinity Fabric |
| Process node | TSMC 3/6nm |
| First availability | Sep 2025est — estimated / secondary-source figure |
| Availability | Shipping |
| Est. purchase price | ~$25,000est — estimated / secondary-source figureAnalyst estimate of average selling price |
CDNA 4, liquid-cooled. Values use AMD's dense BF16/OCP-FP8 and MXFP4 specifications; sparse headline rates are excluded.
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
$8.60/GPU-hrCheapest on-demand · Oracle OCI · observed
1 provider: Oracle OCI. Quotes checked through .
MI355X rental prices from every provider →Value ratios
| $ per dense BF16 PFLOP-hour | $3.44 |
|---|---|
| Dense BF16 TFLOPS per watt | 1.79 |
| Bandwidth bytes per 1,000 BF16 FLOP | 3.20 |
| $ per dense FP8 PFLOP-hour | $1.72 |
| Dense FP8 TFLOPS per watt | 3.57 |
| Bandwidth bytes per 1,000 FP8 FLOP | 1.60 |
| $ per GB of memory per hour | $0.030 |
Prices divide the cheapest on-demand list price; theoretical ratios do not measure application speed.
Sources
Specifications reviewed . Use explicit dense BF16/OCP-FP8 and MXFP4 rates, not the adjacent sparse headline values.
Purchase price: CNBC, citing HSBC ↗ (2025-07-10): “the ASP for MI355 can be USD25k”
Rental price: Oracle OCI published pricing ↗ (BM.GPU.MI355X.8; 8 × MI355X 288GB).