Oracle AI World 2026: the biggest highlights so far
Las Vegas · 25–28 October 2026 · Updated 28 October
The big story is that Oracle wants to move AI from answering questions to actually carrying out business work — using enterprise data, AI agents, cloud infrastructure and applications together.
Here are the announcements I’d put at the top of the list.
1. OpenAI and Oracle are bringing AI directly into enterprise systems
Major announcement
On 27 October, Oracle announced an expanded collaboration with OpenAI to bring ChatGPT, OpenAI models and Codex into Oracle’s enterprise ecosystem. Oracle ASEAN +1
The announced capabilities include:
Oracle AI Workspace: access business information and carry out approved workflows across Oracle Fusion Applications, NetSuite, Oracle AI Database and OCI.
Codex and Oracle App Composer: use AI coding capabilities to build applications, with Oracle handling deployment and operations on OCI.
Oracle AI Security Mesh: use AI-assisted security capabilities to identify and address vulnerabilities across cloud and on-premises environments.
AI tools for smaller businesses: an announced bundle combines ChatGPT Business, Oracle services and cloud credits, subject to terms and eligibility.
Why it matters: This goes beyond putting a chatbot beside an ERP system. The aim is to let users describe what they want to achieve and have AI interact with the underlying business systems to get it done.
For Oracle developers, this could change how custom applications are built and integrated. For administrators, it raises important questions about access controls, auditing, data governance and how AI-generated changes are tested before deployment.
Source: Oracle’s official OpenAI announcement.
2. Google AI chips are coming to Oracle Cloud Infrastructure
More choice beyond NVIDIA GPUs
Oracle and Google Cloud announced that Google’s Tensor Processing Units (TPUs) will become available through OCI, with availability planned for 2027 and a multi-year rollout. Oracle says this will make OCI the first hyperscale cloud outside Google to offer Google TPUs. oracle.com
The significance is that organisations will have more choice when selecting hardware for demanding AI training and inference workloads.
Oracle is expanding its AI infrastructure beyond a single accelerator ecosystem. The wider announcements also point towards additional accelerator options, including OpenAI’s in-house chip and Cerebras hardware. Quartz +1
Why it matters: More hardware options could improve performance, capacity and cost for particular workloads. However, actual benefits will depend on availability, software compatibility, pricing and the workload being run.
Official announcement: Google TPUs on OCI.
3. Oracle is making AI agents a core part of Fusion Applications
Oracle’s 26 October announcements introduce new ways for AI to execute work across its business applications, including ERP, HCM, supply chain, sales and service. oracle.com +1
The two key capabilities are:
Fusion Companion: users state a business objective, and the system coordinates the relevant applications, transactions, workflows, approvals and AI agents.
Fusion Agentic Home: an AI-native workspace where users can delegate work and track its progress.
Oracle is also introducing Fusion Claw-powered applications for complex supply-chain tasks, including allocating scarce materials, planning shipment loads and analysing procurement spending. These applications can evaluate options and prepare or execute authorised actions within organisational controls. oracle.com
Why it matters: The shift is from AI that merely recommends a course of action to AI that can perform multi-step work, subject to permissions and approvals.
For Oracle HCM, ERP and SCM customers, this could be particularly significant. The practical test will be how reliably agents handle exceptions, explain decisions and maintain a clear audit trail.
4. AI could significantly accelerate Oracle Fusion implementations
Oracle announced the Fusion Setup and Configuration Registry (FSCR), a public Git repository that makes Fusion Applications configuration knowledge machine-readable and executable by AI agents. Oracle ASEAN
The idea is to give AI agents access to maintained Oracle implementation knowledge so they can:
Translate business requirements into implementation plans.
Work through configuration dependencies and sequencing.
Execute configuration tasks.
Test and validate the results for expert review.
This is intended to reduce repetitive manual work during implementations and upgrades.
Why it matters: Oracle consultants and implementation partners could spend less time repeating standard configuration tasks and more time on architecture, business process design, integrations and complex customer requirements.
It is also worth watching from a developer perspective: Oracle is treating implementation expertise as reusable, executable skills rather than relying entirely on static documentation.
Official source: Fusion Setup and Configuration Registry.
5. Oracle Autonomous AI Lakehouse gets a major upgrade
Oracle announced enhancements to its Autonomous AI Lakehouse, including a business semantic layer, expanded Apache Iceberg support, AI-assisted data discovery and centralised security controls. oracle.com
The most interesting additions are:
Business Semantic Layer: gives AI agents consistent definitions for business metrics, relationships, calculations and KPIs.
AI Data Catalog: helps discover and understand data across Oracle and non-Oracle systems.
Expanded Apache Iceberg support: improves the ability to work with open table formats and data in object storage.
Business Simulation Engine: brings Essbase capabilities into the lakehouse for planning, forecasting and what-if analysis.
Data Studio: provides a unified workspace for data preparation, integration and analysis using SQL, Python or natural language.
Oracle says the enhanced platform is designed to work across OCI, AWS, Microsoft Azure, Google Cloud and selected on-premises or dedicated-cloud environments. Some capabilities are planned for future availability rather than immediate general availability. oracle.com
Why it matters: AI is only as useful as its understanding of the data behind it. Consistent business definitions, access controls and lineage become increasingly important when AI agents are making decisions or executing workflows.
For Oracle database professionals, this is one of the most technically relevant announcements.
Official source: Autonomous AI Lakehouse enhancements.
6. Oracle commits US$1 billion to its partner ecosystem
On 23 October, Oracle announced a US$1 billion commitment over four years to support its partner ecosystem. Oracle América Latina
The investment is intended to help partners develop AI-powered solutions, build technical skills, pursue cloud migrations and bring offerings to market through Oracle Marketplace.
For New Zealand, this is worth watching. Local Oracle partners may have more opportunities to build AI services, modernise customer systems and develop industry-specific solutions.
For NZOUG, it is also a useful topic for discussion with sponsors and the Oracle partner community.
Official source: Oracle’s partner investment announcement.
What these announcements mean for Oracle professionals
DBAs and database architects
Focus on AI Database capabilities, vector search and retrieval, open data formats, security, and how AI workloads affect database performance and governance.
Developers and architects
Watch the OpenAI integration, Codex, application-building workflows and agent orchestration. These could change how custom applications are developed and maintained.
Cloud and infrastructure teams
Evaluate OCI’s expanding accelerator options, AI infrastructure costs, data residency, security boundaries and the operational implications of deploying AI at scale.
Fusion, ERP and HCM specialists
Pay attention to AI-driven business processes, agent permissions, automated configuration and the distinction between AI recommendations and authorised execution.
My overall take
The most important development at Oracle AI World 2026 isn’t a single product release. It’s the convergence of three things:
Enterprise data that AI can access and understand.
AI agents that can reason about business objectives and perform tasks.
Cloud infrastructure capable of running increasingly demanding AI workloads.
Oracle is positioning itself to provide all three, while integrating with external AI providers and other cloud ecosystems.
The biggest question now is execution: how quickly these features become generally available, how well they work in real production environments, and whether the cost and governance requirements make them practical for customers.
For the NZOUG community, I’d prioritise the OpenAI integration, the Autonomous AI Lakehouse updates and the Fusion agent announcements for a technical debrief or follow-up webinar.
For the complete, up-to-date announcement list, see the official Oracle AI World website: Oracle AI World 2026 — announcements and keynotes. Oracle’s event page also provides keynote replays, which are useful for checking the technical detail behind the headlines.
Comments