Is Amazon Web Services's B200 SXM good for LLM Inference?

AIMC's fit-scoring methodology rates the B200 SXM as excellent fit for LLM Inference (score 100/100). Reference pricing at Amazon Web Services: $14.24/hr. AIMC has no commercial relationship with Amazon Web Services.

What does LLM Inference on Amazon Web Services's B200 SXM cost?

At the median reference price of $14.24/hr: 1 day costs $341.80, 1 week costs $2392.59, 1 month costs $10253.95 per GPU. Pricing sourced from publicly listed catalogs.

Hyperscaler Reference/AWS/B200 SXM/for LLM Inference

Datacenter GPUs · Workload Fit · Reference Pricing

AWS B200 SXM
for LLM Inference

Fit analysis and reference pricing for the B200 SXM at Amazon Web Services, evaluated for LLM Inference.

Workload Fit

100/100

Excellent Fit

Reference Price

$14.24/hr

1 publicly listed offer

Why This Fit?

AIMC's fit-scoring methodology evaluates the B200 SXM's suitability for LLM Inference based on published specifications.

·Datacenter class is well-suited for LLM Inference
·180 GB VRAM provides ample headroom (15.0x the minimum)
·2250 FP16 TFLOPS substantially exceeds the 50 TFLOPS threshold

Estimated Cost for LLM Inference

Per-GPU cost at the median reference price ($14.24/hr). Actual costs may vary based on region, commitment terms, and additional services not reflected in baseline pricing.

1 hour

$14.24

per GPU

1 day (24h)

$341.80

per GPU

1 week (168h)

$2,392.59

per GPU

1 month (720h)

$10,253.95

per GPU

← Back to GPU

AWS B200 SXM

All workloads + alternatives

Other GPUs at AWS

All AWS GPUs for LLM Inference

Workload comparison

AWS Catalog

All Amazon Web Services GPUs

Full reference catalog

Legal Notice — Independent Market Intelligence

AIMC has no commercial relationship or data-sharing agreement with Amazon Web Services. Pricing data is sourced from publicly listed catalogs. "Amazon Web Services", "AWS", and related marks are trademarks of their respective owners. Fit-scoring is AIMC's independent analysis based on published GPU specifications.

Hyperscaler Reference/AWS/B200 SXM/for LLM Inference

Datacenter GPUs · Workload Fit · Reference Pricing

AWS B200 SXM
for LLM Inference

Fit analysis and reference pricing for the B200 SXM at Amazon Web Services, evaluated for LLM Inference.

Workload Fit

100/100

Excellent Fit

Reference Price

$14.24/hr

1 publicly listed offer

Why This Fit?

AIMC's fit-scoring methodology evaluates the B200 SXM's suitability for LLM Inference based on published specifications.

·Datacenter class is well-suited for LLM Inference
·180 GB VRAM provides ample headroom (15.0x the minimum)
·2250 FP16 TFLOPS substantially exceeds the 50 TFLOPS threshold

Estimated Cost for LLM Inference

Per-GPU cost at the median reference price ($14.24/hr). Actual costs may vary based on region, commitment terms, and additional services not reflected in baseline pricing.

1 hour

$14.24

per GPU

1 day (24h)

$341.80

per GPU

1 week (168h)

$2,392.59

per GPU

1 month (720h)

$10,253.95

per GPU