
- Architecture: NVIDIA Hopper
- GPU memory: 141 GB HBM3e
- Cooling: Passive (server airflow)

Product image for illustration; configuration may differ.
AMD's fastest Instinct GPU: 288 GB HBM3E and 8 TB/s per GPU, sold as an eight-GPU platform in a server. Priced on quote.
| Cooling | Air or direct liquid cooling, depending on platform |
|---|---|
| GPU memory | 288 GB HBM3E |
| Peak FP8 | 5 PFLOPS (10.1 with sparsity) |
| Form factor | OAM module on an 8-GPU UBB platform |
| Peak FP16 / BF16 matrix | 2.5 PFLOPS |
| Architecture | AMD CDNA 4 |
| Board power | 1400 W |
| Peak FP64 | 78.6 TFLOPS |
| Memory bandwidth | 8 TB/s |
| Peak MXFP4 / MXFP6 | 10.1 PFLOPS |
| Compute units | 256 (16,384 stream processors) |
| Software | AMD ROCm |
AMD Instinct MI355X 288 GB (OAM, 8-GPU platform) is built to order and priced per project (configuration, delivery, installation and support).
Pick the model and quantisation you plan to run. We'll show whether it fits in memory, how much room is left for context, and the throughput to expect. Estimates come from memory bandwidth and are replaced by our own bench figures where we have them.
Machine
Quantisation
Comfortable. 70 GB spare — room for 128k context, a second small model, or a RAG index in memory. Expect roughly 3.8 tok/s single-stream.
Machines that run this configuration
Weights 42.8 GB at 4.85 bits/param · KV cache 5.0 GB at 16k (fp16, approximate) · 10 GB reserved for OS and runtime. Speed is single-stream decode, memory-bound; prompt processing and batching behave differently. Or WhatsApp an engineer on +44 7515 356 146.
The MI355X is the top of AMD's MI350 series, running at up to 1400 W for the most performance per GPU. Eight of them in one platform give 2.3 TB of HBM3E, enough to serve or fine-tune the largest open models in a single server.
OAM modules are not sold as loose cards: they come as an eight-GPU AMD Instinct platform (UBB) inside a server such as the Dell PowerEdge XE9785, XE9785L or XE9680, or another OEM platform. We quote the whole system.
It runs on AMD ROCm, the open software stack, with day-one support for PyTorch and the main open models.
Fitted in a Workstation or Dell server,. Before it ships we check it against your chassis (power, airflow, slot layout and platform support), burn-in test it and install the AMD ROCm drivers and software stack. Warranty claims and advance replacements are handled by us, and support comes from our own UK engineers.
Priced on quote: tell us what you need and we will confirm price, availability and lead time.
Systems we build carry the warranty stated on your quote: three years return-to-base as standard. Parts carry their manufacturer's warranty, and we help you claim it.
Prices are shown ex VAT. VAT (20%) is added in your basket and on your quote. You can put your own purchase order number on a quote.
UK delivery is £35.00 ex VAT, shown before you pay. For high-value systems we may book a delivery slot or ask for a signature.
Business customers can return unopened stock items within 30 days of delivery (ask us for a returns number first). If you buy as a consumer, you can cancel within 14 days of receiving the goods. Faulty goods are always covered.



