On-demand reads, writes and storage for a DynamoDB table.
DynamoDB write units cost about five times a read unit, so batching, condensing items and avoiding hot-key rewrites saves real money. For steady traffic compare with a relational option on the managed database cost calculator.
The DynamoDB Cost Calculator estimates your monthly on-demand table bill from three inputs: read requests per month, write requests per month, and stored data in GB. It multiplies each request volume by AWS's per-million on-demand request rate and adds the per-GB monthly storage charge, then sums them into a single projected total. The main drivers are request counts and the read-to-write ratio, since writes are priced higher than reads on demand; storage usually contributes a smaller share until your dataset grows large. Entering realistic monthly volumes gives you a bill estimate before you provision anything.
The key trade-off to watch is the on-demand versus provisioned decision. On-demand pricing charges per request with no capacity planning, which is convenient for spiky or unpredictable traffic but can cost more at steady, high volume. Use this calculator to spot when your request counts climb high enough that reserved or provisioned capacity would likely be cheaper, and to see whether reads, writes, or storage dominate your bill so you can focus optimization where it actually moves the number.
On-demand charges per million read request units (about $0.25) and per million write request units (about $1.25) — writes cost roughly 5x reads. Storage is about $0.25 per GB-month after a 25GB free tier.
On-demand is simplest and great for spiky or unpredictable traffic. If your read/write rate is steady, provisioned capacity (optionally with auto-scaling and reserved capacity) can be substantially cheaper per request.