Data-Center AI Accelerator Specs, Performance & Price (July 2026)
Comparison of major data-center AI accelerators (NVIDIA / AMD / Google / Intel / AWS): memory, bandwidth, compute, power and cloud rental price. Specs from official documents, retrieved 22 July 2026.
| Vendor | Chip | Memory | Bandwidth (TB/s) | BF16 (TFLOPS) | FP8 (TFLOPS) | TDP (W) | Year | Cloud rental ($/GPU/hr) |
|---|---|---|---|---|---|---|---|---|
| NVIDIA | H100 SXM | 80 GB HBM3 | 3.35 | 989 | 1979 | 700 | 2022 | $2.99 |
| NVIDIA | H200 | 141 GB HBM3e | 4.8 | 989 | 1979 | 700 | 2024 | $4.39 |
| NVIDIA | B200 | 192 GB HBM3e | 8.0 | 2250 | 4500 | 1000 | 2024 | $5.89 |
| NVIDIA | B300 (Blackwell Ultra) | 288 GB HBM3e | N/A | N/A | N/A | 1400 | 2026 | $7.39 |
| AMD | MI300X | 192 GB HBM3 | 5.3 | 1307 | 2615 | 750 | 2023 | N/A |
| AMD | MI325X | 256 GB HBM3e | 6.0 | 1307 | 2615 | 1000 | 2024 | N/A |
| AMD | MI355X | 288 GB HBM3e | N/A | N/A | 5000 | 1400 | 2025 | N/A |
| TPU v5p | 95 GB HBM2e | 2.77 | 459 | N/A | N/A | 2023 | N/A | |
| TPU v6e (Trillium) | 32 GB HBM | 1.64 | 918 | N/A | N/A | 2024 | N/A | |
| TPU v7 (Ironwood) | 192 GB HBM | 7.37 | N/A | 4600 | N/A | 2025 | N/A | |
| Intel | Gaudi 3 | 128 GB HBM2e | 3.67 | 1678 | 1678 | 900 | 2024 | N/A |
| AWS | Trainium2 | 96 GB | 2.9 | 655 | 1287 | N/A | 2024 | N/A |
Method & sources
Specs (memory / bandwidth / BF16 and FP8 compute / TDP / year) are from each vendor's official datasheets and product pages; BF16 and FP8 are dense (no sparsity) figures. Fields the vendor has not disclosed are marked "N/A" — no estimates. Performance is expressed as official BF16 dense TFLOPS; MLPerf submissions are system-level throughput that cannot be cleanly normalised per chip, so they are not used here (see mlcommons.org for cross-reference). Prices are RunPod on-demand cloud rental (USD per GPU per hour, retrieved 22 July 2026), from a single source for consistent comparison; chips RunPod does not list are "N/A", and non-public enterprise purchase prices are excluded. Chips without an official datasheet (e.g. Huawei Ascend, third-party/leaked data only) and unreleased chips are excluded. AI chip specs and cloud prices change frequently — verify on the official pages.
Source: https://www.nvidia.com/en-us/data-center/h200/
Retrieved: 2026-07-22