Expectations expectation medium confidence

Expected: Databricks Removes Partner-Powered AI Toggle on November 1

Workspaces are expected to keep their existing enabled or disabled value when the control disappears, making the period before removal the clearest time for administrators to verify and document the setting.

Edited by Tyronne Panaino

Databricks expects to remove the partner-powered AI features setting from all workspaces and its Settings API on November 1, 2026. The company says the removal will leave each workspace's existing value in place: an enabled setting should remain enabled, and a disabled setting should remain disabled. Administrators are therefore facing a control change, not an announced automatic activation or deactivation of the underlying feature set.

The change matters to teams that govern which external partner models can support Databricks AI features. According to the partner-powered AI documentation, the setting covers designated external-model providers including Azure OpenAI, OpenAI on Databricks and Anthropic on Databricks. A preserved value can continue to affect the AI capabilities available in a workspace even after the familiar account and workspace controls disappear.

What is known

Databricks' upcoming-change record and the product documentation give the same November 1 removal date. They say that account and workspace toggles are already disabled where the setting is currently enabled. In that state, administrators cannot turn the value off through the user interface before removal.

The pre-removal path differs for a workspace where the setting is disabled. Databricks says that toggle remains available, so an administrator can still enable it in the interface. The company also says the Settings API can continue to enable or disable the value until November 1. That creates a short administrative window in which teams can inspect the current state, decide whether it matches policy and preserve an auditable record of any change.

After November 1, Databricks says customers that need to change the value should contact their account team. The documentation does not describe a replacement self-service control or publish response times, approval criteria or an automated migration route for post-removal changes. Teams that rely on repeatable infrastructure or policy automation should treat that missing operating detail as an unresolved dependency rather than assuming the old API will remain callable.

Default behavior also varies by account configuration. Databricks says partner-powered AI features are enabled by default for new accounts unless an account is configured for compliance security profiles, where the default is disabled. That statement describes the initial default; it does not override the separate promise that existing workspace values will be retained when the control is removed.

What administrators should verify

The first practical check is the value for every affected workspace, including whether account-level and workspace-level views agree. Teams should also identify which enabled products actually depend on partner-powered AI. Databricks lists capabilities that become unavailable when the setting is disabled, so the same policy decision can have different operational effects across workspaces.

A second check is ownership. If a workspace must change after November 1, the documented route moves from a self-service toggle or API call to an account-team request. Administrators should record who can open that request, what internal approval it needs and which workflows would be affected while a change is pending. This is especially important where deployment automation currently assumes that the Settings API can enforce the value directly.

This run did not access a customer account, inspect a live workspace or call the Settings API. It therefore does not establish that the current UI behaves identically in every region, that the API will stop at a particular hour on November 1, or how quickly Databricks will process later change requests.

What would confirm it

The first confirmation checkpoint is the November 1 product change itself: the toggles and API control should no longer be available while sampled workspaces retain their previous values. Updated Databricks documentation or API responses should then show the post-removal management path without implying that the partner-powered features themselves were removed.

Stronger evidence would include independently observed account and workspace behavior across an enabled workspace, a disabled workspace and an account using compliance security profiles. A documented account-team change completed after removal would clarify the new operating process, but no such post-change evidence can exist yet.

Status

Expectation. Databricks has documented the planned removal date and value-preservation behavior, but the November 1 transition has not occurred. Internal confidence is medium because both sources are official Databricks pages and this run did not independently test the administrative controls.

Sources

Update note: Last reviewed 2026-10-06. We will revise this post after the setting is removed or if Databricks publishes a replacement control, a more specific cutoff time or a documented post-removal change process.

Sources

Drafted with AI assistance from source briefs; reviewed for citation completeness and label accuracy.

More Expectations coverage