BTC
ETH
USDT
XMR
LTC
GPU comparison · Ada Lovelace vs Hopper
NVIDIA RTX 4090 vs H100.
A GeForce card against the Hopper data-center GPU. The H100 SXM has 3.3× the memory (80 GB of HBM3 against 24 GB of GDDR6X), 3.3× the bandwidth, about 6× the dense FP8 and BF16 throughput and NVLink between GPUs. The RTX 4090 costs less than a sixth as much: $269 against $1,779 per GPU per month.
- 24 vs 80 GBGPU memory
- 1.01 vs 3.35 TB/smemory bandwidth
- $269 vs $1,779per GPU, per month
RTX 4090 vs H100: which one to rent.
Choose the RTX 4090 if
- Your model fits in 24 GB per card: about 18B parameters in FP8 or 34B in 4-bit by our sizing rule.
- You generate images, render, or prototype before scaling up, where 24 GB per card is enough.
- Price comes first: four RTX 4090 ($1,076 a month) cost less than one H100.
Choose the H100 if
- Your model needs more than 24 GB per GPU: one H100 holds about 61B parameters in FP8, or a 70B model in 4-bit (about 46 GB).
- You train or fine-tune across GPUs: NVLink 4 links H100s at 900 GB/s per GPU, while GeForce cards talk over PCIe.
- You need ECC memory, FP64 for scientific code, or the highest throughput per GPU.
RTX 4090 vs H100 specs, side by side.
Figures from NVIDIA’s published specifications, per GPU. Tensor figures are peak dense throughput.
| Per GPU | NVIDIA RTX 4090 | NVIDIA H100 | H100 / RTX 4090 |
|---|---|---|---|
| Architecture | Ada Lovelace, TSMC 4N | Hopper, TSMC 4N | — |
| Form factor | PCIe Gen4 graphics card | SXM5 module | — |
| GPU memory | 24 GB GDDR6X, 384-bit | 80 GB HBM3, 5,120-bit | 3.3× |
| Memory bandwidth | 1,008 GB/s | 3.35 TB/s | 3.3× |
| GPU-to-GPU link | None (PCIe Gen4) | NVLink 4, 900 GB/s per GPU | — |
| FP4 tensor (dense) | Not supported | Not supported | — |
| FP8 tensor (dense) | 330 TFLOPS | 1,979 TFLOPS | 6× |
| FP16 / BF16 tensor (dense) | 165 TFLOPS | 989 TFLOPS | 6× |
| FP64 | — | 34 / 67 TFLOPS | — |
| Max power | 450 W | Up to 700 W, configurable | — |
| GPUs per server | 1, 2, 4 | 1, 2, 4, 8 | — |
| Per GPU in our servers | 12 vCPU · 64 GB RAM · 1 TB | 20 vCPU · 200 GB RAM · 2 TB | — |
Sources: NVIDIA GeForce RTX 4090 · Ada whitepaper · NVIDIA H100. Full sheets: RTX 4090 · H100
RTX 4090 vs H100 rental price.
At CryptGPU, the RTX 4090 rents for $269 and the H100 for $1,779 per GPU per month, 30% below the market median of public on-demand prices, paid in crypto with no KYC. Hourly figures are the monthly price divided by 730 hours.
| CryptGPU | NVIDIA RTX 4090 | NVIDIA H100 | H100 / RTX 4090 |
|---|---|---|---|
| Price per GPU, per month | $269 | $1,779 | 6.6× |
| Equivalent per GPU-hour | $0.37 | $2.44 | — |
| Per GB of GPU memory, per month | $11.21 | $22.24 | 1.98× |
| Per PFLOPS of dense FP8, per month | $815 | $899 | 1.1× |
| Largest server | 4× · $1,076/mo | 8× · $14,232/mo | — |
| Market median, per GPU | $387 | $2,548 | — |
| Below the median | −30% | −30% | — |
Same price per GPU at every server size, no hourly metering. How we compare prices · Configure an RTX 4090 server · Configure a H100 server
Which models fit on each.
Smallest server that holds each model, by precision. Rule of thumb with headroom for the KV cache, 92% of GPU memory usable.
| Model (total parameters) | RTX 4090 · FP8 | H100 · FP8 | RTX 4090 · 4-bit | H100 · 4-bit |
|---|---|---|---|---|
| Qwen3.5-9B | 1× | 1× | 1× | 1× |
| Gemma 4 31B | 2× | 1× | 1× | 1× |
| Llama 3.3 70B | 4× | 2× | 4× | 1× |
| gpt-oss-120b | — | 2× | 4× | 2× |
| DeepSeek-V4-Flash | — | 8× | — | 4× |
| Qwen3.5-397B-A17B | — | 8× | — | 4× |
| DeepSeek-R1 (671B) | — | — | — | 8× |
| Kimi K2.6 (1T) | — | — | — | — |
FP8 ≈ 1.2 bytes and 4-bit ≈ 0.65 bytes per parameter. — = larger than the biggest server of that GPU. Size another model · How much VRAM does an LLM need?
RTX 4090 vs H100 FAQ.
More on each GPU: NVIDIA RTX 4090 · NVIDIA H100.
Can an RTX 4090 replace an H100?
For models that fit in 24 GB and for image generation, often yes, at a fraction of the price. For larger models, multi-GPU training, ECC memory or FP64 work, the H100 is the right tool.
How many RTX 4090 match the memory of one H100?
Four RTX 4090, the largest GeForce server, hold 96 GB (about 88 GB usable) for $1,076 a month, against 80 GB for $1,779 on one H100. The four cards share the work over PCIe, which is slower than NVLink.
Which is faster for LLM inference?
Per GPU, the H100: 3.3× the memory bandwidth, which bounds token generation, and about 6× the dense FP8 throughput, which speeds up prompt processing. Real speed depends on the model, the precision and the engine; we do not publish benchmarks.
Other GPU comparisons.
Rent the RTX 4090 or the H100.
Dedicated servers with 1 to 8 GPUs, one monthly price, paid in crypto. Online in under 10 minutes, no KYC.