GPU MODELS / A30
A30 cloud GPU pricing
11 listings across 7 providers. Cheapest published rate: $0.080/GPU/hr at SpotGPUs. Sorted by per-GPU hourly price. Every row links to the provider's own published price.
WHAT IS THE A30
The A30 is the quiet value part of the Ampere line: HBM2 bandwidth (933 GB/s) at 165 W, with MIG support for slicing into four isolated instances. For memory-bandwidth-sensitive inference it often beats GDDR6 cards with higher list compute.
TENSOR COMPUTE
165 TFLOPS FP16 tensor dense (330 sparse)
INTERCONNECT
PCIe 4.0; NVLink bridge 200 GB/s
WHAT IT IS GOOD AT
Bandwidth-sensitive inference
HBM2 at this price is the draw.
Multi-tenant serving
Four-way MIG at 6 GB each.
WHO SHOULD NOT RENT THE A30
Jobs needing more than 24 GB per instance.
Raw FP16 throughput per dollar against newer Ada parts at similar rates.
THE A30 MARKET, BY THE NUMBERS
7 providers publish 11 A30 configurations on the panel today. Published per-GPU hourly rates run from $0.080 (SpotGPUs) to $1.23 (Exoscale), a 15x spread between the cheapest and the most expensive published rate for the same chip. The median listing sits at $0.74/GPU/hr. 9 rows were reported in stock by the provider at last observation; the rest are published catalog rates. All rates are on-demand - committed-use and reserved pricing is excluded.
A30 PRICE TREND, DAILY PANEL SNAPSHOTS
Across every A30 row on the panel, the daily floor moved from $0.350 to $0.080/GPU/hr over 11 days of tracking (-77.1%). The median continuously-tracked listing held flat at $0.35. Snapshots run daily; the full history is public in the repo.
A30 PROVIDER NOTES
SpotGPUs · 1 config · 1 in stockSets the panel floor for this model.
from $0.080/hrMatPool · 1 config · 1 region · 1 in stock1 of 1 configs reported in stock at last check.
from $0.326/hrSeeweb · 1 config · 1 region · 1 in stock1 of 1 configs reported in stock at last check.
from $0.331/hrExoscale · 4 configs · 1 region · 4 in stock4 of 4 configs reported in stock at last check.
from $0.740/hrE2E Networks · 1 config · 1 region · 1 in stock1 of 1 configs reported in stock at last check.
from $0.939/hrInHosted · 1 config · 1 region · 1 in stock1 of 1 configs reported in stock at last check.
from $0.939/hrBUYING A30: WHAT THE PANEL SAYS
At the current floor, one A30 running around the clock costs about $58 a month (730 hours of on-demand arithmetic). Node shape matters: the single-GPU floor is $0.080/GPU/hr against $0.350/GPU/hr on an 8-GPU node (+337.5%). Per-GPU floors by config size: 1x $0.08 · 2x $0.90 · 3x $0.79 · 4x $0.74 · 8x $0.35. Compare the alternatives below before committing - the same budget often buys more than one chip class.
A30 VS THE ALTERNATIVES
Same memory size, 55% more bandwidth from HBM2, slightly more compute. For LLM serving the A30 is usually the better chip per dollar.
Compare live rates: A30 vs A10
A30 PRICING FAQ
What is the cheapest A30 cloud GPU right now?
The cheapest published A30 rate on the panel is $0.080 per GPU-hour at SpotGPUs (A30 x 1, Not published). Every price links to the provider's own published rate.
How much does a A30 cost per hour?
Across 7 providers tracked today, published per-GPU hourly rates for A30 run from $0.08 to $1.23, with a median of $0.74. Committed-use and reserved rates are excluded; everything shown is on-demand.
How many providers rent A30 GPUs?
7 providers publish 11 distinct A30 configurations on Compute Cafe. 9 of those rows were reported in stock by the provider's live endpoint at last observation.
Where is A30 cheapest?
By published per-GPU hourly rate, the cheapest geography for A30 today is china at $0.33/hr (MatPool). The spread across 3 tracked geographies runs to $0.94/hr in india.
Are A30 prices going up or down?
Over the 11 days Compute Cafe has tracked the panel, the A30 floor is falling: $0.350 to $0.080/GPU/hr (-77.1%). The median continuously-tracked listing moved from $0.35 to $0.35 (+0.0%). Daily snapshots are public in the repo, so the trend is auditable.
What does a A30 cost per month?
At the current floor of $0.080/GPU/hr, a A30 running 24/7 costs about $58 per month (730 hours of on-demand arithmetic on the cheapest published rate). Committed-use contracts can price lower; the panel tracks on-demand rates only.
Is a full 8-GPU A30 node cheaper than a single card?
On today's panel: the single-GPU floor is $0.080/GPU/hr and the 8-GPU node floor is $0.350/GPU/hr - more expensive per GPU on a full node (+337.5%). Check the table for the exact configs behind each floor.
Why pick an A30?
When inference is bandwidth-bound and the panel prices it below GDDR6 alternatives - 933 GB/s of HBM2 at 165 W is hard to match in its class.
See also alternatives to the A30 · cheapest GPU-hour per model · inference chips · training chips · by location