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.
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
verifiedCheapest MI300X confirmed in stock right now
Limited- 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
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.
| Provider | Fri11 | Sat12 | Sun13 | Mon14 | Tue15 | Wed16 | Thu17 | Week |
|---|---|---|---|---|---|---|---|---|
| Hot Aisle | 0% | |||||||
| Runpod | 0% |
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.
Price History & Comparison
Full MI300X price history (6 months) →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
| Architecture | CDNA 3 (AMD) |
| GPU Memory | 192 GB HBM3 |
| Memory Bandwidth | 5.3 TB/s |
| FP16 Performance | 1,307 TFLOPS |
| TDP | 750W |
| Interconnect | Infinity Fabric (896 GB/s) |
| Release Year | 2024 |
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.
Normalized from live on-demand rows by GPU count.
Provider count is pulled from current MI300X listings.
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.