Oracle & IBM Power Present: Modernize for AI resilience and security while optimizing costs and existing infrastructure - ILOG SA
Based on the transcript provided, here are the forward-looking statements extracted and summarized:
Explicit Guidance:
- No financial guidance (revenue, earnings, margins, capex, etc.) or macroeconomic outlook was provided in this technical webinar.
- Product Release Timing: Oracle Database 26 AI on IBM Power and AIX is expected to be released in the first half of 2027.
- Long-term Collaboration: Management expects the strategic collaboration between IBM and Oracle to continue beyond the end of 2026 AI release and through 2031 and beyond.
Business-Segment Outlook:
- Oracle 26 AI on IBM Power: A major release with over 300 new capabilities focusing on AI-native features, security, and data processing. This release follows an extended remediation period to address vulnerabilities detected by foundational AI models.
- Platform Support: Full feature parity is planned; all 26 AI capabilities available on other Oracle platforms (including cloud) will be available on IBM Power 11 and Power 10 with AIX 7.2 and 7.3, except for Exadata-native storage indexing.
- Legacy Support: The OpenSSL encryption libraries adopted in 26 AI will be backported to Oracle 19c, making quantum-resistant encryption available on older releases.
- Agent Capabilities: "Private Agent Factory" and "Select AI" equivalents will be made available on-premises via the DBMS Cloud package and MCP servers deployed on AIX, extending cloud-native AI agent capabilities to on-premise environments.
- Operational Efficiency (IBM Power 11): Agentic AI is projected to enable 15x faster detection and resolution of capacity constraints compared to traditional methods, a 60% faster learning curve for new developers, and a 98% success rate for business-to-business sales recommendations using IBM Spire adapters.
- System Availability: IBM Power 11 targets six nines (99.9999%) availability, limiting unplanned downtime to 31 seconds per year.
Market and Competitive Outlook:
- Unified Data Platform: Market trend toward consolidating data platforms to reduce sprawl; 26 AI is positioned as a converged platform supporting relational, JSON, vector, and graph data types via SQL without requiring separate systems.
- Data Access Trends: Increased demand for direct access to external data formats; 26 AI introduces support for Apache Iceberg (read/write to object storage) and federated queries against non-Oracle systems (e.g., SQL Server, Teradata) to enable real-time AI without ETL data movement.
- AI Workload Shift: Enterprise focus shifting from general AI speed to secure, agentic AI workflows requiring native database agents to process unstructured data and enforce guardrails via API constraints.
Macro / Industry Outlook:
- Quantum Security Threats: Acknowledgment of the "harvest now, decrypt later" threat where adversaries store encrypted data for future decryption via quantum computing.
- Regulatory Environment: Increasing regulatory focus on Post-Quantum Cryptography (PQC) and cyber resiliency (e.g., DORA regulation) for critical infrastructure and government procurement.
- AI Vulnerability Landscape: Recognition that foundational models have accelerated vulnerability detection, necessitating a shift in engineering focus from feature development to security hardening and vulnerability remediation.
Strategic Initiatives:
- 26 AI Release & Remediation: Completion of an extensive 8-month remediation period in late 2025/early 2026 to fix bugs found by AI tools before the 2027 launch.
- Compiler Upgrade: Replacement of the previous XLC compiler with a new open-source XLC compiler to improve long-term sustainability and support for Oracle 26 AI on Power.
- Security Architecture: Implementation of IBM Power Cyber Vault (response time <1 minute to ransomware) and immutable, air-gapped backups (ZDLRA) to ensure cyber resiliency.
- Integration with IBM Accelerators: Full utilization of SIMD capabilities and MMA (Matrix Math Accelerator) vector processing on Power 11 for AI workloads.
- Open Source Contribution: IBM contributing to OpenSSL and ONNX runtime projects with Power-optimized encryption and AI runtime capabilities.
- Event Schedule: Full disclosure of release notes and platform support scheduled for the Oracle AI World conference in Las Vegas (following the webinar).
Risks and Headwinds to Outlook:
- Release Delays: The 26 AI release cycle has been extended due to the unexpected volume of vulnerabilities identified by AI-driven testing, delaying the AIX/Power launch relative to other platforms (e.g., Linux/Windows).
- Adoption Complexity: Customers face a need to upgrade infrastructure (to Power 11), operating systems (AIX 7.2/7.3), and apply security patches (CVEs) prior to full 26 AI utilization.
- External Security Threats: Ongoing risks of AI-based threats, bypassing guardrails, and quantum decryption capabilities persist as industry-wide headwinds.
Management Tone on the Future:
- Confidence: High confidence in the comprehensive feature parity between IBM Power and other Oracle platforms; no capabilities are being withheld for the Power stack.
- Commitment: Satya Sharma (IBM Fellow) described 26 AI on Power as a "personal project" and highlighted an "extraordinary" and "long-term" commitment to the partnership.
- Urgency: Strong emphasis on the necessity of upgrading to Power 11 and 26 AI to mitigate "systematic risks" and leverage AI speed, though the release timeline is framed as a necessary trade-off for quality and security.
- Collaboration: Tone reflects deep, historic integration (40 years) with executive engagement and dedicated resources (new hardware, personnel) driving the roadmap.
Miscellaneous:
- Hardware Prerequisites: Spire cards are not required to run Oracle 26 AI on IBM Power.
- Demo Availability: Recordings and slide decks will be distributed later the same day; follow-up planning calls and validation programs are available.
- Support Scope: 26 AI will support both Power 10 and Power 11 generations.
- Company
- ILOG SA
- Keywords
- ILOG SA