Oracle Database 23ai Third-Party Support: Oracle's Cloud-First Release, Your On-Premise Alternative

Oracle Database 23ai is Oracle's latest major release, launched in 2023 with AI Vector Search, True Cache, and property graph SQL at the centre of its marketing narrative. It is also Oracle's most explicitly cloud-first release in the history of the product. On Oracle Cloud Infrastructure, Database 23ai is the default offering across Base Database Service, Exadata Cloud Service, and Autonomous Database. On-premise, Oracle 23ai is available as a perpetual licence — but Oracle's account teams are making it difficult to miss the implicit message: 23ai is engineered for OCI, and on-premise deployments are a path Oracle wishes fewer customers would choose.

For the customers who did acquire Oracle Database 23ai perpetual licences — or who are evaluating it as an alternative to extending their Oracle 19c estate — the support economics are identical to every other Oracle Database version: Oracle's Software Update Licence & Support (SULS) runs at 22% of net licence value annually, escalates with Oracle's unilateral right to increase fees up to 3% per year, and funds a Sustaining Engineering tier that patches catastrophic issues and nothing more. Third-party support for Oracle Database 23ai provides full coverage at 62–65% less than Oracle SULS — removing the subscription premium Oracle uses to make on-premise look expensive relative to OCI.

Oracle's 23ai Strategy — What the Cloud-First Architecture Means for You

Oracle Database 23ai introduced several architectural features that are either cloud-exclusive or require OCI-native services to function fully. AI Vector Search — Oracle's flagship 23ai capability, which enables vector embedding storage and cosine similarity queries within the Oracle Database — is available on-premise in the Enterprise Edition perpetual licence. True Cache, Oracle's read-only cache tier positioned as a managed PaaS service on OCI, has limited on-premise availability. The JSON Duality Views feature, marketed as the "best of relational and document" storage, is fully available on-premise but is most aggressively marketed in the context of Oracle REST Data Services and APEX Cloud.

The commercial implication is straightforward. Oracle wants enterprises evaluating Database 23ai to find that the full feature set is available only on OCI, while on-premise 23ai delivers a subset of capabilities at the same SULS pricing. This creates artificial pricing pressure on on-premise perpetual licence holders — the same playbook used with Oracle Database 19c at Premier Support end and with Oracle Java licensing restructuring in 2023. Third-party support removes Oracle from the commercial equation entirely: you run your 23ai perpetual licence on-premise, maintain full database functionality, and eliminate Oracle SULS.

Oracle Database Version Matrix — TPS Eligibility (2026)

The Application Certification Gap — Why Most Enterprises Cannot Upgrade to 23ai Yet

The most significant practical barrier to Oracle Database 23ai adoption in enterprise environments is application certification. Oracle's major application stacks — EBS, PeopleSoft, JD Edwards, Siebel, Hyperion EPM — each require explicit Oracle certification for each major Database version. Oracle typically certifies new Database versions against its application stack 12–18 months after the Database GA date. For Oracle EBS 12.2, the most widely deployed EBS version, the certification roadmap for Database 23ai was not published as of Q1 2026.

SAP on Oracle presents an identical problem from a different direction. SAP does not certify Oracle Database 23ai for SAP production systems — including SAP ECC, BW, and S/4HANA — as of March 2026. SAP's Oracle Database certification is a separate process from Oracle's own application certification, and SAP's certification cycles for major Oracle releases have historically run 18–24 months behind the Oracle GA date. For the large cohort of SAP-on-Oracle customers, 23ai is not a viable near-term upgrade path regardless of Oracle's support deadline pressure. Oracle Database 19c TPS remains the correct answer for these customers during the certification gap.

The customers who are genuine near-term 23ai candidates are those running Oracle Database as a standalone RDBMS (not as the backend for Oracle or SAP applications), particularly development teams evaluating AI Vector Search for new workloads. These customers are also the cohort Oracle targets most aggressively with OCI Autonomous Database — the fully managed 23ai offering that eliminates DBA overhead at the cost of perpetual licence independence. Our Oracle cloud migration guide analyses the full economics of this transition.

What Third-Party Support Covers for Oracle Database 23ai

GoVendorFree's Oracle Database TPS covers the full Oracle Database 23ai stack under a single support contract. Coverage is not contingent on Oracle's support lifecycle decisions — you receive full supported cover regardless of where Oracle classifies 23ai in its Premier / Sustaining Engineering matrix. Key coverage areas include:

• Multitenant architecture (CDB/PDB): Container database and pluggable database management, PDB cloning, PDB relocation, and CDB fleet management — the core 23ai deployment model for most on-premise installations
• AI Vector Search: Support for the 23ai vector embedding infrastructure, including VECTOR data type, DBMS_VECTOR package, and integration with third-party embedding models running locally
• RAC and Grid Infrastructure: Full RAC 23ai support — SCAN listeners, clusterware (23.x), ASM disk groups, RAC One Node, and Oracle Clusterware
• Data Guard and RMAN: Standby database management, log shipping, switchover/failover procedures, Active Data Guard read replicas, and RMAN backup infrastructure
• Partitioning and Advanced Security: Partition pruning optimisation, interval partitioning, Transparent Data Encryption (TDE), Virtual Private Database (VPD), and Unified Auditing
• Performance Diagnostics and Tuning: AWR-equivalent analysis, execution plan management, SQL Plan Management (SPM), adaptive statistics — without the Oracle Diagnostics and Tuning Pack licence requirement
• Oracle True Cache (on-premise): Support for the limited on-premise True Cache configuration available in Oracle 23ai EE
• JSON Duality Views and Graph SQL: Coverage for JSON Relational Duality Views, property graph SQL syntax (SQL/PGQ), and ORDS integration on-premise
• Security advisory: CVE analysis and compensating controls for Oracle 23ai security vulnerabilities — providing materially better security coverage than Oracle Sustaining Engineering once 23ai enters that phase

Oracle LMS Audit Risk for 23ai Customers

Oracle 23ai introduces new licence metric complexities that the Oracle Licence Management Services (LMS) team is already using in audit conversations. The Named User Plus (NUP) vs. Processor metric choice for 23ai environments with AI Vector Search workloads is not clearly documented in Oracle's master agreement — Oracle's LMS teams have begun treating AI Vector Search queries as Processor-metric triggers in environments where NUP was originally chosen for the database licence. This is a manufactured audit risk, not a legitimate licence interpretation, but it requires active management.

Second, Oracle's 23ai True Cache option, when used in cloud-connected configurations, can trigger Oracle Cloud Services licence entitlements under Oracle's Universal Credit model — pulling on-premise environments into OCI consumption tracking. This is a deliberately opaque commercial mechanism. Our Oracle Audit Defence Playbook covers the specific defences against 23ai-era LMS audit tactics. Third-party support includes standing audit defence as a standard service — Oracle cannot audit as a commercial consequence of having moved to TPS.

Primary Cohort: Banking and Large Enterprise Oracle Estates

The Oracle Database 23ai TPS opportunity is most pronounced in three enterprise segments. First, large banks and financial institutions that run Oracle Database as the tier-1 RDBMS for core banking, trading systems, and regulatory reporting — typically Oracle EE with RAC and Data Guard on dedicated Exadata or commodity hardware. These institutions have large SULS invoices (often £500K–£3M annually for the database layer alone), zero appetite for OCI core banking migration, and regulatory frameworks (PRA, ECB SSM, FINMA) that require explicit change management for any cloud migration affecting tier-1 systems.

Second, NHS trusts and healthcare organisations running Oracle Database as the backend for patient record systems, clinical analytics, and Oracle Health (Cerner) deployments. Oracle Health's push to migrate on-premise Cerner environments to Oracle Cloud is creating simultaneous support cost pressure and migration cost pressure — TPS addresses the cost side while preserving on-premise optionality during procurement negotiations.

Third, Oracle Database customers who acquired 23ai perpetual licences specifically to avoid SAP's Oracle 19c certification gap — only to find that SAP does not certify 23ai either. These customers are in a temporary holding pattern that typically lasts 18–36 months while waiting for SAP certification. TPS on Oracle 23ai is the correct commercial response during this period.

The Extended Support Premium Trap — 23ai vs. 19c Economics

For Oracle Database 19c customers evaluating an upgrade to 23ai specifically to avoid Oracle's Extended Support premium (10–30% above standard SULS pricing), the economics typically do not support the upgrade. Oracle Extended Support for Database 19c ends in April 2027. The upgrade project cost for a 19c→23ai migration — including application compatibility testing, RAC/ASM reconfiguration, and re-certification of custom code against 23ai internals — typically runs £80K–£450K for a mid-to-large Oracle estate. Third-party support on Oracle 19c eliminates the Extended Support premium entirely, costs nothing in implementation, and delivers the same 62–65% saving as TPS on 23ai.

The only scenario where upgrading to 23ai is financially justified before moving to TPS is where an organisation has a specific requirement for AI Vector Search in a production workload, or where Oracle EBS / PeopleSoft certification for 23ai has been published and the business case for application stack consolidation is compelling. In all other scenarios, TPS on Oracle 19c is the higher-ROI decision through the certification gap period.

Four-Profile Oracle Database 23ai TPS Cost Model

The following cost model applies to Oracle Database 23ai Enterprise Edition with standard Processor licence metric. Savings are calculated against Oracle SULS at 22% of net licence value per annum, before any Extended Support surcharge.

Transitioning Oracle Database 23ai to TPS — Five Steps

1. Licence inventory and NUP/Processor metric verification: Confirm all Oracle Database 23ai licences, editions (SE2 vs. EE), and metric choices before transition — this is the foundation of audit defence posture
2. Options and packs review: Identify all Oracle-licenced options (Partitioning, Advanced Security, OLAP, Spatial) and diagnostic packs — TPS contracts are scoped to match the licenced feature set
3. Oracle notice and support termination: Submit formal Oracle support termination notice (30-day requirement under Oracle CSI terms) — GoVendorFree manages this process on your behalf
4. TPS contract activation: Service Level Agreement active within 24 hours of Oracle support termination; 15-minute response SLA for P1 issues from day one
5. Ongoing support and security advisory: Quarterly security advisory, patch release notes, and LMS audit defence briefings — maintained throughout the TPS engagement

Related Oracle TPS Resources

Oracle Database 23ai does not exist in isolation. The most common enterprise scenarios involve Database 23ai alongside Oracle application layers, Oracle Middleware, or Oracle options — each with its own TPS economics:

• Oracle Database 19c TPS — the current largest Oracle DB TPS cohort; Premier Support ended April 2024
• Oracle EBS Third-Party Support — Oracle's flagship ERP application; combined DB + application TPS delivers the highest total saving
• Oracle Fusion Middleware TPS — WebLogic, SOA Suite, OSB, and OID often share the same support contract as Oracle Database
• Oracle Exadata Support Alternatives — Exadata hardware support is separable from Oracle Database SULS; combined TPS strategy available
• Oracle Audit Defence Playbook — covers 23ai LMS audit vectors, processor metric disputes, and True Cache licence risk