watsonx.data Integration Pricing: IBM Resource Unit Cost, DataStage Plans and Databand

IBM does not print a price for watsonx.data integration on its pricing page. Its public cloud catalog API does: the Essentials plan is $26.50 per Resource Unit, effective 1 August 2026, and IBM's own plan description recommends starting at a minimum of 200 Resource Units per month. That works out to $5,300 a month, or $63,600 a year, before any negotiated discount.

We pulled those figures from IBM's global catalog on 20 August 2026. Everything below is reproducible: the endpoint is public, needs no API key, and you can run the same commands yourself in under a minute. That matters, because the vendor page this replaces ends at a "Request a quote" form.

How much does watsonx.data integration cost?

There are two plans in IBM's catalog for the service, and only one of them costs money.

The pricing record is a "Granular Tier" model with a charge unit quantity of 1, so $26.50 really is the price of one Resource Unit and not of a bundle. That distinction is not pedantic. The DataStage records further down quote $96.90 against a charge unit quantity of 50, and anyone reading that field carelessly would overstate the unit cost by a factor of fifty.

The number that turns a unit price into a budget is IBM's own guidance, which sits in the Essentials plan description: "We recommend getting started with a minimum quantity of 200 Resource Units per month to get value out of watsonx.data integration capabilities." Take IBM at its word and the entry point is $5,300 per month. List price is a starting position rather than a final number, and enterprise agreements move it, but it is a far better anchor than walking into a call with nothing.

What is a Resource Unit in watsonx.data integration?

A Resource Unit is a shared consumption currency that every capability in the platform draws from. The Essentials plan description names them in order: data observability, real-time streaming, bulk and batch, ETL and ELT, unstructured data integration, and data replication, all reachable from one interface.

The consequence is easy to miss at signature and impossible to miss at renewal. You are not buying six products with six meters. You are buying one meter that six workloads consume. If you bought the platform mainly for data observability and your batch ETL volume triples next quarter, the observability capability you actually cared about did not change, but your bill did. Consumption pricing rewards a team that can forecast its own workload and punishes one that cannot.

This is a genuinely different shape from per-seat or per-table pricing, and it is worth modeling before you commit. Ask which operations consume Resource Units, at what rate, and what a typical month looks like for a customer of your size. Get the answer in writing.

Where does IBM actually publish the price?

In the IBM Cloud global catalog. It is public, unauthenticated, and it is the single most useful thing we can hand you here, because it works for other IBM services too.

Find the service, then read its plans, then read the pricing on a plan:

BASE=https://globalcatalog.cloud.ibm.com/api/v1

# 1. find the service
curl -s "$BASE?q=watsonx-data-integration"

# 2. list its plans (service id from step 1)
curl -s "$BASE/4d4fd270-fe8f-11ef-82ad-af5f37d5a455/*?complete=true"

# 3. read the price on the Essentials plan
curl -s "$BASE/49814b94-a7c5-4b6c-9cc6-5a0c02a42a2e/pricing"

Three fields decide whether you read the answer correctly. charge_unit_quantity is the divisor, and skipping it is the classic mistake. amounts is a list by country, so take the entry where country is USA rather than the first one in the array. And effective_from tells you how current the record is, which is how we know these are August 2026 prices rather than something stale.

One honest limit: the catalog is the IBM Cloud SaaS list price. Self-managed software, perpetual licenses and anything sold through an IBM Business Partner are priced separately and are not in this API.

Is there a free tier for watsonx.data integration?

Yes. The catalog lists a Trial plan of type "free", and IBM's product pages offer a 30-day free trial of the full capability set. That is enough to load a real pipeline and watch what it consumes, which is the only reliable way to size a Resource Unit commitment. Use the trial as a metering experiment, not just a feature tour: instrument something representative, run it for a week, and extrapolate.

How much does IBM DataStage cost?

DataStage sits under the same platform and, unusually, its list prices are fully exposed in the same catalog. All the figures below carry an effective date of 1 August 2026.

Two notes so you can trust the table. The $1.938 comes from dividing the Standard and Enterprise figure of $96.90 by its charge unit quantity of 50; the independent check is that the Anywhere plans quote the same unit directly at $1.94 for a single Capacity Unit-Hour, and the two agree to the cent. And the API does not declare a billing period for the Instance metric, so we will not tell you whether $8,760 is monthly or annual. It is a question worth asking early, because the two answers differ by more than an order of magnitude.

How much does IBM Databand cost?

There is no separate Databand price any more, because there is no separate Databand. IBM acquired Databand.ai in 2022 and has since folded the product into watsonx.data integration as its data observability component, metered in the same Resource Units as everything else. The domain databand.ai now redirects to an IBM page that does not use the word Databand once.

So the answer to "how much does Databand cost" is the answer to the Resource Unit question above, weighted by how much observability work you run against the shared meter. We walk through the redirect chain, the support lifecycle records that still carry the old product IDs, and what the GitHub and PyPI history shows about the open source SDK, on our page covering IBM Databand pricing and alternatives.

Is watsonx.data integration the same as watsonx.data intelligence?

No, and the names are close enough that shortlists get built against the wrong product.

watsonx.data integration is the movement and reliability platform: ETL and ELT, streaming, replication, and data observability. watsonx.data intelligence is the governance and metadata side, where IBM's catalog, knowledge catalog and the lineage technology from its Manta acquisition ended up. If your requirement is a business glossary, stewardship workflow or enterprise lineage, intelligence is your product line. If your requirement is pipelines that run and data that arrives on time, integration is. Our comparison of metadata management tools covers the governance half of that split across the whole market.

How do you keep a Resource Unit bill from drifting?

Consumption pricing fails quietly. Nobody approves an overage; it accumulates. Three habits keep it honest.

• Meter before you commit. Run the trial against a workload that looks like production, not a toy pipeline, and record actual Resource Unit consumption for at least a week.
• Attribute consumption to a workload. If six capabilities share one meter, you need to know which one moved. Ask IBM what breakdown the platform reports before you sign, not after your first surprising invoice.
• Alert on the trend, not the total. A monthly invoice tells you what happened four weeks ago. Watching the run rate and putting a real-time alert on the budget turns a quarterly finance problem back into a Tuesday engineering problem.

What should you ask IBM before you sign?

Five questions, in the order that saves the most money:

1. How many Resource Units does a workload like ours consume in a typical month, and what drives that number up?
2. Is the Instance charge on the DataStage plans billed monthly or annually?
3. What discount applies at our committed volume, and what happens if we exceed or undershoot it?
4. Which capabilities consume Resource Units at different rates, and can we see per-capability consumption reporting?
5. Given the Databand brand no longer appears in the product, what is the committed roadmap for the data observability component specifically?

That last one is not a rhetorical dig. When a capability you are buying for stops being marketed under its own name, its roadmap belongs to the platform rather than to the feature, and it is fair to ask what that means for the thing you actually need.

Where a smaller tool fits instead

If the reason you are pricing an integration platform is that you need to know what breaks when a column changes, and which tables went stale overnight, this is a large purchase for a narrow problem. Datatrail connects to Snowflake, BigQuery, Databricks or Redshift read-only and builds column-level lineage from your query history and dbt manifest, then runs downstream impact analysis and freshness and schema-change monitoring on that same graph. Nothing is installed in your pipelines, nothing is written back to your data, and the price is on the pricing page at $99, $299 or $799 a month.

That is not a replacement for an ETL platform, and we would not pretend otherwise. If you genuinely need bulk movement, streaming and replication, you need a platform and IBM sells a serious one. If you needed observability and lineage and got quoted a platform, it is worth pricing the narrow tool before the broad one. Our roundups of data observability tools and data observability pricing cover the rest of the field, published prices and quote-only vendors alike.