Zeotap launched Open CDP, positioning it as the first fully composable customer data platform that runs on open-source compute inside a customer's own lakehouse. Zeotap describes the new product as "sovereign by design."

The launch shifts the company's CDP offering from a packaged vendor stack toward an architecture where compute, storage, and identity resolution live on infrastructure the customer already owns.

What does "fully composable" mean?

Zeotap describes Open CDP as the first CDP that breaks every function — identity resolution, segmentation, activation, audience sync — into independent modules a customer can deploy, replace, or skip. In a traditional packaged CDP, those capabilities ship together inside a single vendor stack. A composable CDP exposes each capability as a separate service a team can wire into its own data workflows.

The framing targets martech teams that already run pipelines on lakehouse platforms and want to keep customer data in place rather than copying it to a vendor-managed warehouse.

Why does the lakehouse matter?

A lakehouse combines the structured querying of a data warehouse with the low-cost, object-storage scale of a data lake. Running a CDP on a customer's lakehouse means identity graphs, event streams, and audience tables stay in the same environment where analytics, machine learning, and reverse-ETL jobs already run. That removes the duplicated storage costs and the lag introduced when an external CDP ingests a nightly extract.

For practitioners, the practical question is which compute engines Open CDP supports and which activation endpoints ship at launch. Zeotap does not enumerate those in the announcement.

What does "sovereign by design" imply?

"Sovereign by design" frames the product around data residency, vendor lock-in, and regulatory control. Because compute and storage remain on infrastructure the customer owns, the vendor does not hold a copy of the underlying customer data in its own tenancy. That distinction matters for organizations operating under GDPR, India's DPDP Act, or sectoral rules in financial services and healthcare, where regulators ask where identity data resides and which legal entity processes it.

Where does this sit in the CDP category?

The product's launch lands as more enterprise buyers explicitly ask vendors about exit costs, data egress fees, and the ability to swap identity or activation modules without re-platforming. Open CDP's positioning answers those questions in principle — the data stays on customer infrastructure, and modules can be swapped independently — but the announcement does not yet specify version numbers, GA dates, pricing tiers, or integration partners.

What changes for practitioners?

For a martech leader evaluating Open CDP, three questions now sit at the top of any evaluation:

• Does the product run on the specific lakehouse engine already deployed, or only on a subset?
• What activation endpoints ship at general availability, and which remain on a roadmap?
• How does pricing scale — by event volume, identity count, modules deployed, or compute consumed?

The headline claim of "first fully composable, sovereign CDP" remains a positioning statement until the vendor publishes versions, dates, and integration specifics practitioners can test against existing deployments.