Guide4 sections2 min read
The short answer
An NVIDIA H200 is rented per GPU-hour, and the rate depends on the variant, the region and the moment. On EverXCloud the live rate is the best offer on the exchange, shown in the table below, excluding tax.
What you get with an H200
The H200 is a Hopper GPU with 141 GB of HBM3e memory and 4,800 GB/s of memory bandwidth, against 80 GB and 3,350 GB/s for the H100 SXM. Compute throughput is the same as the H100 SXM: 989 TFLOPS in BF16 and 1,979 TFLOPS in FP8, at up to 700 W.
The extra memory lets larger models or longer contexts fit on one GPU, which matters most for inference of large language models. The H200 NVL variant targets air-cooled servers.
What moves the hourly rate
On an exchange, the rate moves with supply and demand: more offers lower the best price, more demand raises it. Region, GPU count per machine, storage and network included in the configuration also change the price of an offer.
Compare the same variant, memory size and form factor, in the same region. A headline hourly rate rarely describes the full session cost, so add preparation time, storage and any separately billed resources.
Pay the price you choose
If the best offer is above your budget, place a bid at the price you are willing to pay per GPU-hour. The bid is matched when an offer meets it.
During pre-launch, prices shown on EverXCloud are indicative and become rentable on launch day. The table above is read from the live exchange each time the page is generated.
Live prices on EverXCloud
| GPU | Memory | From, per GPU-hour |
|---|---|---|
| NVIDIA H200 141GB | 141 GB | $4.19 |
| NVIDIA H200 NVL 141GB | 141 GB | $3.86 |
| NVIDIA H100 SXM 80GB | 80 GB | $2.88 |
Best live offer on the exchange, excluding tax. Prices are indicative during pre-launch.
Frequently asked questions
- How much does an H200 cost per hour?
- The current rate is the best live offer on the EverXCloud exchange, shown on this page and on the H200 page, per GPU-hour and excluding tax.
- Is the H200 worth it over the H100?
- When a model does not fit in 80 GB, or when memory bandwidth limits inference, the H200's 141 GB and 4,800 GB/s can reduce the number of GPUs needed. For compute-bound work that fits in 80 GB, the H100 SXM offers the same throughput.
- Are the prices shown final?
- No. During pre-launch, EverXCloud prices are indicative. They are shown per GPU-hour, excluding tax, and become rentable on launch day.