AMD Instinct MI355X 8-GPU platform with direct liquid cooling

AMD Instinct MI350 Series GPU accelerator (OAM module)

Important information:

Price incl. VAT — no import fees to pay on delivery
Product imported, inspected and sold by a French company.
Warranty and support in French included.

8-GPU AMD Instinct MI355X AI Server

Request a quote

Hurry up! Only 0 left

10 customers are viewing this product

Delivery

Standard delivery in 7 days

Quantity:

Price incl. VAT, no import fees
Brand new, 1-year warranty, after-sales service in France
Tracked delivery to Austria and all 27 EU countries, no customs fees

Read our FAQ

New product

All products sold on this site are brand new, with no import fees and, above all, a 1-year warranty from the date of purchase.

Description

AMD's alternative for large-scale AI: 8 Instinct MI355X GPUs and 2.3 TB of HBM3e memory, with the open ROCm software stack.

Key features

8 AMD Instinct MI355X GPUs (CDNA 4 architecture) with 288 GB of HBM3e each, for a total of about 2.3 TB of GPU memory
8 TB/s of memory bandwidth per GPU
AMD Infinity Fabric interconnect between the GPUs
FP4 and FP6 compute for inference
Direct liquid cooling (about 1,400 W per GPU)
2 AMD EPYC processors, ROCm software (PyTorch, vLLM…)

Use cases

Inference of very large models (288 GB of memory per GPU means models don't have to be split), training, generative AI, high-performance computing.

How to order

The server is built by an AMD partner manufacturer chosen based on availability. An air-cooled version (AMD Instinct MI350X) is available on request. The final configuration is validated with you in the quote.

On quotation: describe your project using the form at the bottom of the page (target models and use cases, number of servers, available power and cooling, desired lead time). We get back to you with a priced proposal and the detailed configuration; the deposit is requested by bank transfer once the quote is approved.

Manufacturer's warranty and support from our team in France.