# Rent an 8-GPU AMD MI355X server for $19,900 per month

> MI355X rental: a bare-metal 8x AMD Instinct MI355X server (288 GB HBM3e per GPU) for $19,900/month, about $3.41 per GPU-hour. ROCm ready, paid in crypto.

Source: https://overgpu.com/gpus/amd-mi355x  
Updated: 2026-10-23  
Publisher: OverGPU (https://overgpu.com/)

**In short:** OverGPU rents the AMD Instinct MI355X by the month: an 8-GPU AMD Instinct MI355X server (8 × 288 GB HBM3e) for $19,900 per month, about $3.41 per GPU-hour. It is in limited stock and delivered in under 10 minutes after payment. Payment is from a USD balance topped up in Bitcoin, Ethereum, Monero, USDT (TRC-20) or Solana, with no setup fee, and the plan can be cancelled before any monthly renewal.

## MI355X configurations and monthly prices at OverGPU

| Configuration | Type | GPU memory | CPU | RAM | Storage | Network | Price per month (USD) | Per GPU-hour | Availability |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| [8× AMD Instinct MI355X](https://overgpu.com/gpus/amd-mi355x) | 8-GPU server | 8 × 288 GB HBM3e | 2× EPYC 9575F · 128 cores | 3 TB DDR5 | 8× 3.84 TB NVMe | 8× 400G RoCE | $19,900 | $3.41 | Limited stock, delivered in under 10 minutes after payment |

Per GPU-hour = monthly price ÷ number of GPUs ÷ 730 hours, shown only to compare with hourly offers: every plan is billed as one flat monthly price.

## AMD Instinct MI355X specifications (vendor-published)

| Specification | Value | Note |
| --- | --- | --- |
| Architecture | CDNA 4 | AMD |
| GPU memory | 288 GB HBM3e |   |
| Memory bandwidth | 8 TB/s |   |
| FP4 | 10,100 TFLOPS | MXFP4 · Dense |
| FP8 | 5,000 TFLOPS | 10,100 TFLOPS with sparsity |
| BF16 / FP16 | 2,500 TFLOPS | 5,000 TFLOPS with sparsity |
| GPU interconnect | Infinity Fabric | full mesh between the 8 GPUs |
| Form factor | OAM module · liquid-cooled |   |
| Max power | 1,400 W |   |

## What the MI355X is best at

- **Large models per GPU:** 288 GB of HBM3e per GPU holds a 70B model in BF16 on a single GPU with room for KV cache.
- **Many replicas per server:** Eight 288 GB GPUs can serve eight independent 70B replicas without tensor parallelism.
- **Full fine-tunes of 70B models:** 2.3 TB of HBM3e holds the full Adam training state of a 70B model, roughly 1.1 to 1.4 TB.
- **ROCm research and HPC:** A bare-metal AMD Instinct server with 128 EPYC cores and 3 TB of DDR5 for scientific code on ROCm.

## When another GPU fits better

- [Rent NVIDIA B300 and HGX B300 servers from $5,190 per month](https://overgpu.com/gpus/nvidia-b300): If your code depends on CUDA-only libraries or you need one GPU rather than eight, the B300 has the same 288 GB per GPU.
- [Rent NVIDIA H200 and HGX H200 servers from $3,090 per month](https://overgpu.com/gpus/nvidia-h200): When 141 GB per GPU is enough and you want the long-established CUDA inference stack, from one GPU at $3,090.
- [Rent NVIDIA B200 and HGX B200 servers from $4,590 per month](https://overgpu.com/gpus/nvidia-b200): For NVLink 5 at 1.8 TB/s between any two GPUs and FP4 support across the CUDA ecosystem.

## Overview

If you want to rent AMD MI355X capacity, we offer it as a bare-metal AMD Instinct server with eight MI355X GPUs for $19,900 per month. That is 2.3 TB of HBM3e in one machine, paid monthly from a prepaid balance topped up in crypto, with a ROCm image ready on delivery. We do not offer a single-GPU MI355X plan: the 8-GPU server is the only configuration, and stock is limited.

## What the MI355X is

The MI355X is AMD's Instinct accelerator on the CDNA 4 architecture. Per AMD's specifications, each GPU has 288 GB of HBM3e at 8 TB/s and delivers 5.0 PFLOPS of dense FP8 and 2.5 PFLOPS of dense BF16. CDNA 4 also adds support for the MXFP4 format, which AMD lists at 10.1 PFLOPS. The GPU comes as an OAM module rated at 1,400 W and is liquid-cooled.

The interconnect works differently from NVIDIA's HGX servers. AMD lists seven Infinity Fabric links of 153.6 GB/s per GPU, about 1.08 TB/s in total, and in an eight-GPU platform each GPU links directly to each of the other seven. Collective operations such as all-reduce, which spread traffic across all peers, can use all seven links at once; a transfer between just two GPUs is limited to one link. HGX B200 and B300 servers, by contrast, give every GPU 1.8 TB/s of NVLink bandwidth to any peer. Our article on [the AMD MI355X for AI](https://overgpu.com/blog/amd-mi355x-for-ai) covers the architecture in more depth.

## An 8-GPU AMD Instinct server, not a single GPU

If you were looking for MI355X cloud instances sold by the GPU, note that we rent this one only as a whole server. You get the full machine: two AMD EPYC 9575F processors with 128 cores in total, 3 TB of DDR5, eight 3.84 TB NVMe drives (about 30 TB) and eight 400G RoCE network adapters for multi-node work. It is bare metal, so there is no hypervisor between your code and the GPUs.

That shape suits teams with enough work to keep eight GPUs busy: serving several large models, fine-tuning, or research code that already runs on ROCm. If you need only one GPU with 288 GB, the [single NVIDIA B300](https://overgpu.com/gpus/nvidia-b300) offers the same memory per GPU on CUDA, at $5,190 per month.

## Memory math: 288 GB per GPU, 2.3 TB per server

Weights take roughly 2 bytes per parameter in BF16 and 1 in FP8, plus KV cache and 10–30% headroom. With 288 GB per GPU:

- **70B model in BF16:** about 140 GB, so it fits on one GPU with well over 100 GB left for KV cache. One server can run eight independent 70B replicas, with no tensor parallelism and no inter-GPU traffic.
- **405B-class model in FP8:** about 405 GB, which fits on two GPUs. You can run up to four copies per server, or one copy spread over all eight GPUs for a much larger KV cache.
- **671B mixture-of-experts model in FP8:** about 671 GB, which needs at least three GPUs and leaves roughly 1.6 TB of the server for KV cache and overhead.
- **Full fine-tuning of a 70B model:** Adam needs roughly 16–20 bytes per parameter before activations, about 1.1–1.4 TB, which fits in the 2.3 TB of HBM3e with roughly 0.9–1.2 TB left for activations.

Our [VRAM calculator](https://overgpu.com/tools/vram-calculator) runs the same math for other models and precisions.

## ROCm software: what runs and what to check

The server comes with root access over SSH and a ROCm image; Debian or Rocky Linux are available on request. PyTorch has official ROCm builds that keep the familiar `torch.cuda` interface, so PyTorch code that sticks to standard operators usually runs without changes, and vLLM supports ROCm for serving. A first check:

```
rocm-smi                                  # GPUs, memory use, temperature
docker run -it --device=/dev/kfd --device=/dev/dri --group-add video \
  --ipc=host --shm-size 16g rocm/pytorch:latest \
  python -c "import torch; print(torch.cuda.device_count())"
```

What needs testing: custom CUDA kernels, which must be ported to HIP, and libraries that ship only CUDA binaries, such as some quantization or attention kernels. Multi-GPU jobs use RCCL, AMD's counterpart to NCCL. Our [ROCm documentation](https://overgpu.com/docs/amd-rocm) covers the details. Because the minimum term is one month, you can validate your stack on real hardware before you commit to anything longer.

## MI355X price per month and per GPU-hour

The 8-GPU MI355X server costs $19,900 per month, about $3.41 per GPU-hour at full use: the monthly price divided by eight GPUs and 730 hours. Within our own catalog, the HGX B300 also carries 2.3 TB of HBM across eight GPUs, for $39,900, and the HGX H200 carries 1.1 TB for $23,900. The MI355X gives the most GPU memory per dollar of our 8-GPU servers; what you trade is the CUDA ecosystem and NVLink's switched fabric. Our comparisons of the [MI355X and the B200](https://overgpu.com/compare/mi355x-vs-b200) and of the [MI355X and the H200](https://overgpu.com/compare/mi355x-vs-h200) weigh those trade-offs.

Against hourly billing, the break-even point is $19,900 divided by the hourly rate you would pay for eight comparable GPUs. If you need the server for most of the month, monthly wins; for a few days of testing, hourly billing costs less where it is available. The price includes 50 TB of outbound traffic per month with no egress fees, a public IPv4 address and an IPv6 /64, DDoS mitigation, 24/7 support from engineers and the 99.9% availability SLA. There is no setup fee.

## Ordering an MI355X server

Order the server on the [deployment page](https://overgpu.com/deploy): choose the configuration, set up the OS image, optional preinstalled software and your login (root password or SSH key), and create a free account or log in in the same form. You then pay the first month from your prepaid balance, which you top up in BTC, ETH, XMR, USDT on TRON or SOL, with +5% bonus credit for every full $100 in one top-up, up to +50% from $1,000; card payments are coming soon. Root SSH access is delivered in under 10 minutes after payment, with the IP and SSH details in your dashboard. Most orders need only an account and a funded balance; orders for several 8-GPU servers may require verification first. Customers with an active server can ask for volume pricing or 6–12 month terms by ticket from the dashboard.

## Frequently asked questions

### How much does it cost to rent an AMD MI355X server?

The 8-GPU MI355X server is $19,900 per month, about $3.41 per GPU-hour at full use. The price is flat, with no setup fee, and includes 50 TB of outbound traffic per month.

### Can I rent a single MI355X GPU?

No. We offer the MI355X only as a bare-metal 8-GPU server. For a single GPU with 288 GB of memory, see the NVIDIA B300.

### Does PyTorch work on the MI355X?

Yes. PyTorch has official ROCm builds that use the same torch.cuda interface, and vLLM supports ROCm. Custom CUDA kernels and CUDA-only libraries need a ROCm port or an alternative.

### Which operating system images are available?

We provide ROCm images for AMD servers, with root access over SSH. Debian or Rocky Linux are available on request.

### Is the MI355X available now?

Stock is limited. The configurations card on this page shows current availability, and the server is delivered in under 10 minutes after payment.

### Is the MI355X server shared with other customers?

No. It is a bare-metal machine with all eight GPUs assigned to you only, and its disks are securely erased between customers.

## Related pages

- [8-GPU server rental: bare-metal NVIDIA HGX and AMD Instinct](https://overgpu.com/products/8-gpu-servers): NVIDIA HGX B300, B200, H200, H100, AMD MI355X or 8x RTX PRO 6000. 50 TB bandwidth, monthly terms, delivered in under 10 minutes.
- [GPU servers for LLM training on dedicated 8-GPU nodes](https://overgpu.com/solutions/llm-training): Bare-metal HGX B300, B200 and H200 nodes with NVLink and 400–800G networking, rented monthly. See the memory math first.
- [GPU servers for LLM inference, running 24/7 at a flat price](https://overgpu.com/solutions/inference): KV cache sizing, FP8 and FP4, and 20–50 TB of bandwidth included. See the sizing math.
- [GPU servers for research and HPC, on fixed monthly budgets](https://overgpu.com/solutions/research-hpc): H100, H200 and MI355X with vendor FP64 figures, CUDA or ROCm, on fixed monthly budgets. Match the GPU to your code.
- [Ubuntu GPU image with CUDA preinstalled](https://overgpu.com/docs/os-images): What our Ubuntu GPU image with CUDA preinstalled contains, how to check driver and toolkit versions, pick 22.04 or 24.04 and keep the driver pinned safely.
- [GPU cost calculator: monthly vs hourly GPU rental](https://overgpu.com/tools/gpu-cost-calculator): Monthly totals, break-even hours and the extra costs to count before you decide.