Intel Gaudi 3
vs
Nvidia H100
The Gaudi 3 is newer, with 60% more memory and 10% more bandwidth. 1 provider lists it, against 56 for the H100.
At a glance
Memory
60% more on Gaudi 3
- Gaudi 3
- 128 GB
- H100
- 80 GB
Memory bandwidth
10% higher on Gaudi 3
- Gaudi 3
- 3.70 TB/s
- H100
- 3.35 TB/s
Cheapest on-demand
Gaudi 3 is not sold on demand
- Gaudi 3
- --
- H100
- $1.80 /GPU/hr
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
| Billing | Gaudi 3 | H100 | Difference |
|---|---|---|---|
| On-demand | No on-demand listings | $3.55 /GPU/hr from $1.80 | |
| Reserved | No reserved listings | $3.31 /GPU/hr from $1.53 (1 mo) | |
| Spot | No spot listings | $2.62 /GPU/hr from $1.29 |
Heads up: A provider's own page may quote a different figure, for example on tax, region or a promotion. A monthly-only plan shows a derived hourly rate. Medians are across providers, one vote each, so a large catalog does not outweigh a small one. Verify before provisioning. More on how we price.
Price history
Gaudi 3 vs H100 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | Gaudi 3 | H100 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 128 GB HBM2e | 80 GB HBM3 94 GB on NVL | 60% more on Gaudi 3 |
| Bandwidth | 3,700 GB/s | 3,350 GB/s 2,000 GB/s on PCIe 3,938 GB/s on NVL | 10% higher on Gaudi 3 |
| Bus width | 5,120-bit | ||
| Compute · dense | |||
| FP8 | 1,678 TFLOPS | 1,979 TFLOPS | 18% higher on H100 |
| FP16 / BF16 | 1,678 TFLOPS | 989.5 TFLOPS | 70% higher on Gaudi 3 |
| INT8 | Not supported | 1,979 TOPS | H100 only |
| FP32 | 67 TFLOPS | ||
| FP64 | Not supported | 34 TFLOPS | H100 only |
| Platform | |||
| Interconnect | RoCE 1,200 GB/s | NVLink 900 GB/s NVLink 600 GB/s on PCIe, NVL | 33% higher on Gaudi 3 |
| Board power | 900 W | 700 W | 29% higher on Gaudi 3 |
| Released | Q2 2024 | Q3 2022 | 2 years newer on Gaudi 3 |
| Precision | |||
| Supported formats | FP8FP16BF16TF32FP32 | FP8FP16BF16TF32FP32FP64INT8 | |
Figures are for each model's headline configuration. Where a form factor carries a different number, it is named under the figure.
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: Gaudi 3 datasheet and H100 datasheet.
Where to rent them
Providers that rent these GPUs, with each one's cheapest rate. No provider we track offers both right now.
| Provider | Gaudi 3 | H100 |
|---|---|---|
Denvr
Canada
|
On request Custom | -- |
HyperAI
Hong Kong
|
-- | $1.80 /GPU/hr On-demand |
GPU.ai
UAE
|
-- | $2.48 /GPU/hr On-demand PCIe |
Hyperstack
UK
|
-- | $2.60 /GPU/hr On-demand PCIe |
Massed Compute
USA
|
-- | $2.73 /GPU/hr On-demand |
Runpod
USA
|
-- | $3.19 /GPU/hr On-demand NVL |
Jarvislabs
India
|
-- | $3.49 /GPU/hr On-demand SXM |
Lyceum
Germany
|
-- | $3.49 /GPU/hr On-demand |
Runcrate
USA
|
-- | $3.55 /GPU/hr On-demand SXM |
1 provider list the Gaudi 3 and 56 the H100. Every listing is on the Gaudi 3 page and the H100 page.
Similar GPUs
Other comparisons in the same class.
-
-
-
Intel Gaudi 2
96 GB HBM2e · 34% less bandwidth than the Gaudi 3
On-demand from $1.03 /GPU/hr
-
Our data for Intel Gaudi 3 was last updated on Sept. 23, 2026, and for Nvidia H100 on Sept. 14, 2026.