UNITED STATES

GPU servers in the USA

Place GPU compute near your US applications and datasets. Compare models, then confirm the location and resources of the selected server.

Models to consider for US deployment

Hopper

H100

80 GB HBM3

Price on requestConfiguration and region to be confirmed
Ampere

A100

80 GB HBM2e

Price on requestConfiguration and region to be confirmed
Ada Lovelace

L40S

48 GB GDDR6 ECC

Price on requestConfiguration and region to be confirmed
Ada Lovelace

RTX 4090

24 GB GDDR6X

Price on requestConfiguration and region to be confirmed

Before choosing a location

01

Data proximity

Identify where your datasets and application backends are stored. Cross-region transfers affect both the workflow and its cost.

02

Network requirements

Specify required bandwidth and connectivity. Measured latency depends on the actual route; geography alone is not a latency guarantee.

03

Recovery location

Choose where recovery copies should be kept and how data will be restored if the primary environment is unavailable.

Questions before you start

Can I select a specific US city?

Save the city in your workload request. A region match alone does not confirm that city; the precise location needs a separate availability check.

Is a global benchmark a US quote?

No. This page only uses offers or benchmark records explicitly scoped to the US region.