H100 vs H200
Datasheet specs side by side, plus what each card actually costs to rent today across every provider we track.
H100
H200
Spec sheet
| H100 | H200 | |
|---|---|---|
| Architecture | Hopper | Hopper |
| VRAM (max variant) | 94 GB | 143 GB |
| FP16 TFLOPS (dense) | 989 | 989 |
| Memory bandwidth | 3350 GB/s | 4800 GB/s |
| TDP | 700 W | 700 W |
| Interconnect | NVLink 900GB/s | NVLink 900GB/s |
| Launched | 2022 | 2023 |
Computed price-performance at today's cheapest on-demand rates: H100 $2.01 per PFLOP-hour vs H200 $2.63. Formula: cheapest $/hr ÷ dense FP16 PFLOPS (1 PFLOP = 1000 TFLOPS). Specs from vendor datasheets (source, source).
What they cost to rent (2026-08-22)
| H100 | H200 | |
|---|---|---|
| Cheapest on-demand $/hr | $1.99 | $2.60 |
| Cheapest spot/community $/hr | $0.94 | $2.51 |
| Cheapest source | Voltage Park | GMI Cloud |
| Providers offering it | 32 | 22 |
Cheapest verified on-demand rate per side from today's snapshot; highlighted = cheaper. Spot/marketplace tiers shown on the individual pages.
FAQ
Is the H100 cheaper to rent than the H200?
H100 is cheaper: the H100 rents from $1.99 per GPU-hour and the H200 from $2.60 on-demand, verified 2026-08-22.
Which has more VRAM, H100 or H200?
The H100 has 94 GB and the H200 has 143 GB. The H200 has more.
Prices render from the live dataset each build. More: the screener · price movers · methodology.