What is the AMD MI300X GPU?

AMD MI300X is a 192 GB high‑bandwidth GPU on the CDNA 3 architecture, designed for large‑scale AI training and inference workloads.

Which cloud offers the cheapest on‑demand MI300X?

As of August 2025, Vultr lists MI300X at $1.85 per GPU‑hour in its Chicago region.

Is Thunder Compute planning to add MI300X?

Thunder Compute is monitoring MI300X availability; today we focus on NVIDIA A100s and H100s while maintaining industry‑leading cost‑performance.

Why compare MI300X to Thunder Compute’s A100?

A100s remain the best bang‑for‑buck for many training and fine‑tuning tasks; showing both helps developers pick the right GPU for their budget.

Back

AMD MI300X Pricing (September 2025): Cheapest High‑Memory GPUs in the Cloud

On‑demand hourly rates for AMD’s 192 GB MI300X across major U.S. providers

Published:

Aug 1, 2025

Last updated:

Sep 16, 2025

Current on‑demand MI300X prices (per GPU)

Provider	SKU / Instance	On‑Demand $/GPU‑hr*	Notes
Vultr	MI300X single‑GPU	$1.85	Chicago region, pay‑as‑you‑go
TensorWave	8 × MI300X bare‑metal	$1.50	$12/hr node ÷ 8 GPUs
RunPod	MI300X 192 GB	$2.99	Community Cloud on‑demand
Sesterce	MI300X 192 GB	$3.30	Hourly marketplace rate
[Vultr]	8 × MI300X cluster	$3.99	$31.92/hr node ÷ 8 GPUs
Oracle Cloud	BM.GPU.MI300X.8	$6.00	8‑GPU bare‑metal ($48/hr)
Azure	ND MI300X v5 (8 × MI300X)	$7.86	$62.85/hr VM ÷ 8 GPUs in West US

*Normalized to a single MI300X, even when only multi‑GPU nodes are offered.

Methodology (why you can trust these numbers)

On‑demand only – no reservations or prepaid contracts.
Same silicon – every figure refers to AMD MI300X (192 GB) SKUs.
Public price lists – pulled September 16, 2025 from each provider’s published pricing pages.
U.S. regions, USD – prices elsewhere vary by 5‑20 %.

Cost‑performance snapshot

Provider / GPU	Two‑hour runtime	Effective cost
Thunder Compute – A100 80 GB	2 × $0.78	$1.56
Vultr – MI300X	2 × $1.85	$3.70
TensorWave – MI300X	2 × $1.50	$3.00
RunPod – MI300X	2 × $2.99	$5.98
Azure – MI300X	2 × $7.86	$15.72

Thunder Compute’s A100 still delivers ~2‑5× more runtime per dollar than most on‑demand MI300X options, making it the cost‑efficiency pick for prototyping and fine‑tuning.

Why this matters for developers

More memory, new math – MI300X packs 192 GB HBM3 and FP8 support, ideal for 70 B+ parameter models.
Price gap vs. H100 – Even the cheapest MI300X ($1.85/hr) costs ~25 % less than the lowest H100 price we tracked in July 2025.
A100 still wins ROI – Unless your model truly needs 192 GB or FP8, an A100 80 GB at $0.78/hr often delivers better value.
Bookmark this page – we refresh the numbers quarterly so you don’t have to.

Takeaways

Vultr and TensorWave have opened the MI300X price war below $2/hr.
RunPod remains the easiest MI300X self‑service experience under $3/hr.
Hyperscalers (Azure, Oracle) still charge 3‑4× more than niche clouds.
Thunder Compute A100s stay the absolute price‑performance leader for most workflows.

Carl Peterson

Try Thunder Compute

Start building AI/ML with the world's cheapest GPUs