Oracle Database@Google Cloud: New Licensing Model

For years, running Oracle on Google Cloud meant falling back to the punitive on-premises core factor rule. The Oracle–Google multicloud partnership changed that — Oracle Database@Google Cloud now runs Oracle-managed infrastructure inside Google data centres under the same favourable counting as OCI. This guide explains the new Oracle Database@Google Cloud licensing model, where the savings are real, and where the lock-in hides.

By Oracle Practice Lead

What Changed: GCP as an Authorized Cloud

The Oracle Database@Google Cloud licensing model exists because of the 2023 Oracle–Google multicloud partnership, which made Google Cloud an Authorized Cloud Environment for Oracle for the first time. Before that, generic Oracle workloads on Google Cloud were treated as unlisted infrastructure and counted under the on-premises core factor rule — the most expensive way to licence Oracle anywhere. The new service flips that: Oracle operates Exadata and Autonomous Database infrastructure physically inside Google data centres, and you consume it as a managed service with a single Google bill.

The practical effect is that Oracle Database@Google Cloud carries the same licensing economics as Oracle Cloud Infrastructure, not the same economics as a self-managed database on a Google Compute Engine VM. That distinction matters: the favourable terms attach to the managed service, not to Oracle running anywhere on GCP. Run your own Oracle binaries on a generic GCE instance and you are back under the core factor rule. This is one of the cloud-specific subtleties the advanced Oracle licensing guide covers across every hyperscaler.

The service itself comes in two flavours that mirror OCI: Autonomous Database, where Oracle handles patching, tuning and high availability automatically, and Exadata-based services for the most demanding workloads. Both run on Oracle-operated hardware sited in Google regions, with networking that places the database physically close to your Google-native application tier to keep latency low. For an enterprise already committed to Google Cloud for analytics or application hosting, that proximity is the real draw — it removes the cross-cloud latency and egress penalties that previously made running Oracle alongside BigQuery or GKE awkward and expensive.

BYOL vs License Included

Like OCI, the service offers two commercial models. License Included bundles the Oracle Database licence into an all-inclusive hourly rate — simplest to procure, and the right choice when you have no spare entitlement. BYOL lets you apply perpetual licences you already own and pay only the lower infrastructure-and-service rate. The gap between them is large: BYOL reduces Autonomous Database compute cost by roughly 76% versus License Included.

A 76% compute saving is only a saving if you already hold the licences and are paying 22% annual support on them anyway. BYOL converts a sunk support cost into deployed value; License Included makes sense when you have no entitlement to redeploy and want to avoid a fresh perpetual purchase.

The decision is identical in shape to the one on AWS and on Oracle's own cloud, which is why it should never be made platform-by-platform. Compare the position here against the Oracle on AWS BYOL rules and the OCI pricing comparison before committing, because the same perpetual licences can usually only be in one place at a time, and the most efficient home is rarely obvious from the per-hour rate alone. The Exadata Cloud Service licensing detail applies directly, since the GCP service runs on the same Exadata infrastructure.

Converting Your Existing Entitlement

BYOL only works if your perpetual entitlement actually covers the service footprint you intend to run. Oracle Database@Google Cloud uses the same conversion as OCI: one Processor licence — or 25 Named User Plus — covers eight ECPUs or two OCPUs. Map your owned entitlement to that ratio before you size the service, not after.

You ownCovers (ECPU)Covers (OCPU)Best for
1 Processor licence8 ECPUs2 OCPUsDirect BYOL mapping
25 Named User Plus8 ECPUs2 OCPUsUser-metric estates
10 Processor licences80 ECPUs20 OCPUsMid-size Autonomous DB

The reconciliation often reveals that an estate has more entitlement than it can use on the new service — classic shelfware — or less than it assumed. Either finding is valuable before you sign. Pair this exercise with a licence harvesting review to surface unused rights you can redeploy, and the broader optimisation-after-migration analysis so the BYOL position reflects your real footprint rather than your historic purchase.

Watch the support arithmetic too. BYOL only converts a sunk cost into value if you are already paying 22% annual support on the licences you bring; if the licences are unsupported, you cannot use them under BYOL without reinstating support first, and Oracle's reinstatement fees can erase much of the apparent saving. Model the all-in position — subscription rate, the support you keep paying, and any reinstatement cost — rather than comparing the BYOL hourly rate against License Included in isolation. A 76% headline compute discount can shrink considerably once the support line is included honestly.

The Multicloud Negotiation Traps

The convenience of a single Google bill comes with concentrated lock-in. Once your database, your application tier and your data gravity all sit inside the Oracle–Google service, the cost and complexity of leaving rise sharply — which weakens your leverage at the next renewal. Treat the initial commitment as the moment of maximum negotiating power, because it is. Push for committed-rate protection, documented exit and data-egress terms, and clarity on how BYOL entitlement is verified, before signing rather than after.

There is also a dual-vendor dependency to weigh. Because the service is jointly delivered, your support and commercial relationship spans both Oracle and Google, and an issue at the boundary — networking, billing reconciliation, or a feature available on OCI but not yet on the GCP variant — can fall into a gap between them. Confirm which party owns each support tier, how billing disputes are resolved, and whether the GCP service has feature or region parity with native OCI before you migrate a production system that depends on it.

Watch, too, for the policy-versus-contract gap that runs through all Oracle cloud arrangements: favourable counting rules published as policy can change, so the terms that matter belong in your agreement. For the full methodology, the Oracle Negotiation Playbook documents the multicloud negotiation framework, and the Oracle vendor intelligence hub tracks the partnership terms as they evolve. If you are evaluating Oracle Database@Google Cloud against AWS, Azure or OCI, request a confidential briefing before you commit — the platform decision and the licensing decision should be made together, not in sequence.

Common Questions

Oracle Database@Google Cloud: FAQ

Is Google Cloud an Authorized Cloud Environment for Oracle?
Yes. Following the Oracle–Google multicloud partnership, Google Cloud became an Authorized Cloud Environment, so standard Oracle BYOL counting rules now apply to qualifying Oracle Database@Google Cloud services. This is a change from the earlier position, where generic Oracle workloads on Google Cloud fell back to the on-premises core factor rule.
How much does BYOL save on Oracle Database@Google Cloud?
Bringing your own licence reduces Autonomous Database compute cost by roughly 76% compared with the License Included rate. The trade-off is that you must already hold — and keep paying 22% annual support on — the underlying perpetual licences. BYOL only wins when you have entitlements you can redeploy rather than re-buy.
How do my existing Oracle licences convert to Google Cloud compute?
Oracle Database@Google Cloud uses the same conversion as OCI: one Processor licence — or 25 Named User Plus — covers eight ECPUs or two OCPUs. Mapping your perpetual entitlement to this ratio before you size the service is the only way to know whether BYOL actually covers your intended footprint or leaves a gap.
Who manages and bills Oracle Database@Google Cloud?
Oracle operates the database infrastructure — patching, high availability, autonomous features — inside Google data centres, while you consume it as a service and receive a single bill through Google Cloud. That single-vendor convenience is real, but it also concentrates lock-in, which is why the commercial terms deserve scrutiny before signing.

Don't Sign an Oracle Multicloud Deal Alone

The initial commitment is the moment of maximum leverage — and the moment lock-in begins. We model the BYOL position and negotiate the terms before you sign.

Request a Confidential Briefing See Our Oracle Case Study

Related guide: Oracle Autonomous Database Licensing: What You Need to Know

Oracle Licensing Intelligence

Monthly briefings on Oracle policy changes, audit tactics, ULA strategy and cloud licensing — from advisors who have been on both sides of the table.