Elastic Pricing: Plans, Costs & What You Actually Pay
Elasticsearch Serverless is billed in VCU-hours rather than by node: ingest from $0.14 per VCU-hour, search from $0.09 and machine learning from $0.07, with storage at $0.047 per GB per month. Hosted and self-managed deployments are priced differently.
Quick answer
Elastic publishes its prices. Per VCU-hour for compute, per GB per month for storage and retention, per GB for egress. Metering is based on uncompressed, normalised, fully enriched data volume, which is much higher than the raw size on the wire..
Free option: Vector profiles receive 50GB free. Workflows include 10,000 free executions and Agent Builder includes 1,000 free executions per month.
Trial: The pricing page links to a free trial of Elastic Cloud; the serverless page does not state a fixed trial length.
Plans at a glance
| Plan | Price | What it is |
|---|---|---|
| Elasticsearch Serverless | from $0.14 per VCU-hour (ingest) | Fully managed, usage-based. Search from $0.09 per VCU-hour, machine learning from $0.07. |
| Storage and retention | from $0.047/GB/month | Charged on retained data; metering uses enriched volume rather than raw bytes. |
| Egress | from $0.05/GB egress | Charged per GB transferred per month. |
| Elastic Cloud Hosted | Resource based | Priced by hardware configuration, cluster size and node count; pay as you go monthly or prepaid. |
| Self-managed | Licence based | Licensing based on number of nodes and used RAM, with Platinum and Enterprise subscription tiers. |
How billing works
The billing metric is: Per VCU-hour for compute, per GB per month for storage and retention, per GB for egress. Metering is based on uncompressed, normalised, fully enriched data volume, which is much higher than the raw size on the wire..
Fees that are easy to miss
- Support is charged as a percentage of the bill: Standard included, Gold 5%, Platinum 10%, Enterprise 15%.
- Elastic Managed LLM: $4.50 per million input tokens and $21 per million output tokens.
- Elastic Inference Service: from $0.08 per million tokens, priced per model.
- Cross-project search: from $0.0023 per GB retained plus $0.05 per GB transferred.
How the vendor frames its pricing
- Serverless removes capacity planning: Elastic manages hardware configuration, cluster size, node count and versions, and scales on search and indexing load.
- The metered volume is enriched data, not raw bytes, so a compression ratio you assumed does not translate into the bill.
- Hosted, serverless and self-managed are three genuinely different cost models for the same product, not three tiers of the same thing.
Before you sign: what actually moves the price
- Estimate on enriched volume. The page warns that metered volumes are much higher than the raw or compressed size on the wire.
- Workflows and Agent Builder have free execution allowances, but the per-execution price after that is low enough that volume matters more than the tier.
- Support is a percentage on top of the whole bill, so Platinum adds 10% to everything, including storage and egress.
- If you need customer-managed keys, traffic filtering or cross-project search, check the feature roadmap: serverless does not include everything hosted does.
Frequently asked questions
How is Elasticsearch Serverless priced?
Per VCU-hour: ingest from $0.14, search from $0.09 and machine learning from $0.07. Storage and retention from $0.047 per GB per month and egress from $0.05 per GB.
What is a VCU?
A search, ingest or machine-learning compute unit that Elastic scales automatically based on load. You are billed per VCU per hour rather than per node.
Is there a free allowance?
Vector profiles receive 50GB free. Workflows include 10,000 free executions per month and Agent Builder includes 1,000.
How much does support cost?
Support is a percentage of the charge: Standard is included, Gold is 5%, Platinum is 10% and Enterprise is 15%.
Why is my bill higher than my data size suggests?
Ingest and retention are metered on uncompressed, normalised, fully enriched data volume, which Elastic states will be much higher than the raw or compressed size on the wire.
Where this data comes from
Primary source: https://www.elastic.co/pricing/serverless-search — retrieved 2026-10-22.
Prices change. Every figure here is a snapshot taken on the retrieval date shown on each page. Confirm with the vendor before signing.