Skip to content
GPU Finder

AMD MI300X cloud pricing and rental comparison

As of , the cheapest single MI300X confirmed in stock is $2.99 per GPU-hour on Hot Aisle. The lowest listed price is $2.39 per GPU-hour on Runpod, but that configuration is currently reported out of capacity. GPU Finder compares 2 providers for this size and refreshes availability hourly. Methodology.

Compare AMD MI300X rental options across 2 cloud providers, including current stock signals for 192GB AMD Instinct accelerators.

Providers:2
Whole-node range:$2.39 — $19.12/hr
Instances:6
Available:1 providers
Free egress:2 providers
Updated:

Structured offer basis: 2 current 1× MI300X listings range from $2.39 to $2.99 USD per physical GPU-hour. Whole-node prices for other GPU counts are not mixed into this range.

Cheapest single MI300X rental

verified

Cheapest MI300X confirmed in stock right now

Limited
$2.99/hr1× listing
on Hot Aisle · us-michigan
Visit Hot Aisle
Is now a good time to rent an MI300X cloud GPU?
near its 6-month high The 4× MI300X on-demand floor is $9.56/hr ($2.39/GPU-hr) — 6-month range $7.97–$9.56/hr for 4× nodes. In short, prices are near their 6-month high, so it may pay to wait or lock in a spot rate.
What is the cheapest MI300X cloud GPU available right now?
Cheapest MI300X confirmed in stock right now: $2.99/GPU-hr on Hot Aisle (1× listing, $2.99/hr).

Prices are the cheapest listed on-demand rates; “in stock” means the provider reported capacity on our last hourly check. How we compute this →

MI300X rental prices by provider

  • 1x MI300X Secure Cloud
    Out0% · 30d
    $2.39/hr
    1× MI300X8 vCPU64 GBFree egressGlobal
  • vm-1x-mi300x
    Limited0% · 30d
    $2.99/hr
    1× MI300X8 vCPU224 GBFree egressus-michigan
  • 2x MI300X Secure Cloud
    Out0% · 30d
    $4.78/hr
    2× MI300X16 vCPU128 GBFree egressGlobal
  • vm-2x-mi300x
    Limited0% · 30d
    $5.98/hr
    2× MI300X26 vCPU448 GBFree egressus-michigan
  • 4x MI300X Secure Cloud
    Out0% · 30d
    $9.56/hr
    4× MI300X32 vCPU256 GBFree egressGlobal
  • 8x MI300X Secure Cloud
    Out0% · 30d
    $19.12/hr
    8× MI300X64 vCPU512 GBFree egressGlobal

Get notified when MI300X drops in price or reliability changes

One email per change, max once a day. We send a confirmation link first; one-click unsubscribe in every email.

MI300X (1× GPU) availability — last 7 days

No provider held 1× MI300X ≥80% available across all 7 days. Polled hourly · 2 providers tracked · 7-day window.

≥80%50–80%<50%No data
View:
Provider
11
12
13
14
15
16
17
Week
Hot Aisle0%
Runpod0%

30-day reliability depth

GPU Finder does not treat availability as a static yes/no flag. For MI300X, current stock badges are paired with a 30-day reliability score based on our hourly stock checks — the share of tracked time each listing was reported available. How we compute this →

6
listings with a 30-day score
100%
of listings on this page have a score
30d
longest stock history on this page

Strongest reliability signals here: Runpod (0%, 30d covered), Hot Aisle (0%, 30d covered).

Caveat: scores stay hidden until a listing has at least 48 tracked hours, and some provider APIs expose coarse capacity levels instead of exact stock counts. Use the score as historical depth beside current availability, not as a guarantee that a GPU will still be allocatable when you click through.

Loading chart...

About the MI300X

The AMD Instinct MI300X is the strongest alternative to NVIDIA's datacenter GPU dominance. Built on the CDNA 3 architecture with 192 GB of HBM3 memory, it offers 2.4x the memory capacity of the H100 at competitive performance levels. This makes it particularly attractive for large-model inference where the entire model needs to fit in GPU memory.

Key Specifications

ArchitectureCDNA 3 (AMD)
GPU Memory192 GB HBM3
Memory Bandwidth5.3 TB/s
FP16 Performance1,307 TFLOPS
TDP750W
InterconnectInfinity Fabric (896 GB/s)
Release Year2024

Cloud Pricing Context

MI300X pricing ranges from roughly $1.71/hr to $6/hr, with most neoclouds at $2-3/hr. It undercuts H100 pricing by 10-30% on a per-GPU-hour basis and offers substantially more memory per dollar. ROCm software maturity has improved significantly through 2025.

Best For

  • Large-model inference where 192 GB VRAM eliminates tensor parallelism
  • Teams seeking NVIDIA alternatives without sacrificing performance
  • Memory-bound workloads such as long-context LLM serving
  • Organizations with ROCm-compatible stacks (PyTorch, JAX)

MI300X pricing at a glance

Fast checks for AMD MI300X rental intent before the full provider table.

Lowest per-GPU-hour
$2.39

Normalized from live on-demand rows by GPU count.

Tracked providers
2

Provider count is pulled from current MI300X listings.

192GB memory-per-dollar
80.3

GB per dollar-hour; compare against H100's 80GB capacity.

At $2.39/GPU-hr, MI300X delivers 80.3 GB per dollar-hour at the current floor. That 192GB footprint is the practical H100 comparison point for memory-bound LLM inference, long-context serving, and workloads where fewer large-memory GPUs simplify deployment.

MI300X cloud pricing FAQ

How much does an AMD MI300X cost in the cloud?

GPU Finder currently tracks AMD MI300X cloud pricing from $2.39/GPU-hr across 2 providers. The table below shows live provider, region, availability, and spot/on-demand rates.

Can I rent MI300X by the hour?

Yes. The MI300X rental listings on this page are hourly cloud GPU rates; use the provider table to compare on-demand and spot options before launching capacity.

Is MI300X cheaper than H100?

Compare MI300X and H100 on memory-per-dollar, not only raw benchmark scores. MI300X has 192GB of HBM3 per GPU versus 80GB on H100, so memory-bound inference can need fewer GPUs even when headline hourly rates are close.