Oracle's Database 26ai goes on-prem, but draws skeptics • The Register

Oracle has launched its Database 26ai Enterprise Edition for Linux x86-64, marking the first general on-premises release of its AI-focused database iteration. However, the extended support timeline for the previous version 19c and concerns about AI lock-in are making many users think twice before upgrading.

Extended support for 19c creates upgrade hesitation

Oracle's decision to extend Premier Support for Database 19c until December 31, 2029, with Extended Support running through December 31, 2032, has created a significant barrier to adoption for the newer 26ai release. Originally slated to end in 2024, this 13-year support window for 19c gives organizations an unusually long runway to avoid upgrading.

"They're not looking at anything soon," said Martin Biggs, vice president and general manager of third-party support specialist Spinnaker. "Vendor application requires a database, and so that database will typically fully support Oracle 19c, and that's going to be the case for six years. It's a pretty stable platform, and people seem pretty happy."

This extended support timeline effectively allows organizations to plan for migration to entirely new applications or alternative databases like PostgreSQL without the pressure of an impending support cutoff.

AI features come with lock-in concerns

The Database 26ai release introduces several AI-focused capabilities, including AI Vector Search, support for JSON Relational Duality, Apache Iceberg table format compatibility, and tooling for building, deploying, and managing AI agents that operate on private data. However, these features come with significant concerns about vendor lock-in.

Mark Smith, CEO of third-party support specialist Support Revolution, warned that Database 26ai is "optimized for Oracle engineered systems" and that costs are likely to increase once customers are locked into the model. "Running the AI content on-premise outside of Oracle kit will almost certainly see an increase in resources needed and a revaluation of machine capabilities and required license metrics," Smith said.

Support portal confusion adds to uncertainty

The confusion around Oracle's release and support schedules has been exacerbated by the company's new support portal, which launched late last year to significant user criticism. The portal has made it difficult for users to understand Oracle's long-term plans for database releases across different platforms.

While Database 26ai is currently available only on Linux x86-64, it remains unclear when it will launch on other operating systems such as Windows. This limited availability, combined with the unclear support timeline for 26ai (Premier Support is slated to end in 2031, with Extended Support presumably available for another three years, though this hasn't been officially announced), creates additional uncertainty for potential adopters.

Market context and user preferences

Despite Oracle's continued dominance in the database market according to DB-Engines rankings, PostgreSQL has emerged as the most popular database among developers, according to Stack Overflow. This preference for open-source alternatives may influence organizations' decisions as they evaluate whether to upgrade to Oracle's AI-focused database.

Oracle's history with database releases shows a pattern of dividing its offerings into "long-term" releases with extended support periods and "innovation" releases with shorter support windows. Database 23ai was positioned as a long-term release but never made it to broadly deployable on-premises platforms, creating skepticism about whether 26ai will follow a different path.

Strategic implications for organizations

The combination of extended 19c support, AI lock-in concerns, and unclear support timelines for 26ai creates a complex decision landscape for organizations. Many appear to be taking a wait-and-see approach, particularly given the stability and familiarity of 19c.

For organizations considering their database strategy, the current situation presents several options: continue with 19c through its extended support period while planning for eventual migration; cautiously evaluate 26ai while monitoring lock-in risks; or explore alternative database solutions like PostgreSQL that may offer more flexibility and lower long-term costs.

As Oracle continues to push its AI initiatives through its database offerings, organizations will need to carefully weigh the benefits of cutting-edge AI capabilities against the risks of increased vendor dependency and potential cost escalations. The extended support for 19c provides breathing room for this evaluation, but also raises questions about Oracle's long-term strategy for database innovation and support.