When AI-powered threats move at scale and speed, defense has to move faster.
AI is changing how attackers discover weaknesses and how quickly they can exploit them. For database owners and security teams, exposure can begin anywhere: an unpatched system, a weak configuration, a compromised credential, or an application connected to sensitive data.
Databases have traditionally operated behind network firewalls, with tightly controlled access through applications and administrative tools. Those boundaries are no longer sufficient as agentic AI and “vibe-coded” applications using a broadly privileged database connection can also expose data without an attacker ever breaking in.
Weak cryptography with insecure algorithms, outdated protocols, and poorly managed keys or certificates can widen exposure. Quantum computing adds a longer-term risk as attackers can capture encrypted traffic today and retain it for future decryption. Customers need to identify these weaknesses across the fleet so that they can strengthen security.
These risks span the entire database estate. Disconnected, database-by-database reviews make it difficult to identify exposure and act consistently. Organizations also need to preserve the ability to recover when an attack, operational mistake, or infrastructure failure disrupts the business.
Oracle’s strategy follows three priorities: Secure at Speed, Secure at Source, and Secure through Resilience. Find and fix weaknesses across the fleet faster. Enforce security where the data resides. Preserve the ability to recover quickly.
Today at Oracle AI World, we are announcing capabilities designed to help organizations assess the security of their database fleet and shorten the time to fix any gaps. New cryptographic assessments in Data Safe help identify weak configurations and guide remediation. Data Subsetting in Data Safe helps teams prepare smaller, masked datasets for faster patch testing. Planned Database Security Central updates extend support for additional PostgreSQL, MySQL, Windows, and Linux/RHEL releases.
To strengthen security at the source, we are expanding Deep Data Security authorization for analytics, APEX, ORDS, and AI agents. We are also introducing ezTLS to simplify network encryption without customer-managed certificates.
Secure at Speed
Security and database teams need to answer practical questions quickly. Which databases are missing updates? Where are sensitive records exposed? Which accounts have excessive privileges? Which systems still allow weak encryption? Before hackers discover these weaknesses, organizations need to understand the gaps and close them quickly.
Oracle Data Safe and Oracle Database Security Central bring security posture, user risk, sensitive-data discovery, monitoring, and policy management for your database fleet into a shared operational view. Centralized assessment and monitoring also help teams identify configuration drift as environments change. Oracle provides deployment choices for centralized database security. Oracle Data Safe is an Oracle-managed cloud service while Oracle Database Security Central is a customer-managed platform for Oracle and non-Oracle databases.
Introducing Cryptographic Posture Assessment in Oracle Data Safe
Data Safe now reports on weak cryptography across eligible Oracle databases, providing configuration details and remediation guidance for encryption at rest, in transit, and in backups. For example, a security administrator can easily identify databases that permit deprecated TLS 1.0 or TLS 1.1 protocols, or certificates expiring within two weeks. Learn more.
The assessment also supports post-quantum transition planning. It distinguishes databases that need release upgrades from those that need configuration changes to use quantum-resistant capabilities. A Cryptographic Bill of Materials, or CBOM, provides the configuration details needed for investigation.
An upcoming release of DBSAT 26.1 also adds cryptographic posture assessment for command-line reviews, alongside updated CIS Benchmark and STIG coverage.
Introducing Data Subsetting in Oracle Data Safe for smaller test datasets
AI is accelerating vulnerability discovery and the release of security fixes. Organizations need to apply those fixes promptly, but testing can become a bottleneck. Full production-sized test copies consume storage, take time to prepare, and increase the compute required for testing. Smaller, representative datasets can reduce that burden while preserving the relationships needed to test application behavior.
With the newly added support for data subsetting in Data Safe, teams can build a much smaller but representative dataset with far fewer rows and their related data. An associated masking policy can replace sensitive values with fictitious ones in the same workflow. This approach can help organizations reduce storage requirements and potentially lower associated compute costs to run the tests. For example, reducing a 2 TB test database to 200 GB means 90% less data to store and process. Additionally, subsetting helps reduce the data exposure risk by limiting the amount of sensitive information used in non-production environments. Learn more.
Automatically bring Exadata databases into Oracle Data Safe security assessments
With the new integration for Exadata Database Service, registering a container database with Data Safe automatically registers its existing and newly created pluggable databases. Data Safe then initiates configuration and user assessments so teams can review findings and prioritize remediation. This integration brings new databases into Data Safe’s assessment coverage as soon as they enter your fleet, reducing the risk that they are overlooked. Learn more.
Expanding coverage in Oracle Database Security Central
A planned release update for Database Security Central is coming soon. The update expands audit support, including before-and-after values, for PostgreSQL 16, 17, and 18, and MySQL 8.4 and 9.7. With this update, organizations can use Database Security Central to collect audit data for Windows Server 2022 and 2025, and Linux/RHEL 10. In addition, the update is expected to support centralized monitoring of end-user activity across Oracle AI Databases secured with Deep Data Security.
The free promotion program for Database Security Central continues through February 28, 2027. To learn more about Security Central, read the announcement blog.
Secure at Source
An application, API, or agent can choose its next query, but the database must decide whether it is allowed. Oracle enforces authorization, encryption, SQL Firewall policies, and privileged-access controls directly in the database. Those controls continue to apply when an agent takes an unexpected path or a new application is introduced. With this secure at source approach, customers can expand how they use data without relying on every application and agent to implement the security rules correctly.
Enforce user’s data permissions across applications and AI Agents
Applications commonly connect through service accounts that can access data for many users. Agents can choose queries, invoke tools, and pursue paths their developers did not anticipate. A misguided agent may stray from the intended workflow while continuing to use valid credentials and broad database privileges.
With Oracle Deep Data Security, applications and agents propagate the end user’s identity, roles, and context through cryptographically signed tokens that the database validates. Declarative SQL policies called data grants define which rows and columns the user can access and which operations they can perform.
Today, we are announcing the expansion of Deep Data Security for analytical applications. New data grant filters let organizations apply additional access restrictions, such as limiting a regional sales manager’s access to records associated with their territory or business unit. This builds upon the earlier announced support for related tables, remote data sources, and advanced SQL analytics.
New integration of Deep Data Security with Oracle APEX and Oracle REST Data Services propagates the end user’s identity and application context to the database for low-code applications, REST services, and supported AI agent workflows. Configurable identity-provider integration helps customers use existing enterprise identities.
Deploy TLS encryption without certificate management
Oracle’s proprietary Native Network Encryption (NNE) is widely used because customers can encrypt database traffic without managing and distributing certificates or wallets.
We are introducing Oracle Database ezTLS to simplify network encryption with standards-based TLS, without requiring customers to manage certificates or wallets. Certificate-authenticated TLS remains the choice when clients or servers must also prove their identity through certificates. ezTLS is available starting with Oracle AI Database 26ai Release Update 23.26.4 and Oracle Database 19c Release Update 19.33. Learn more.
Protect today’s traffic against tomorrow’s quantum threat
An attacker does not need a quantum computer today to create a quantum-era problem. They can record encrypted traffic now and wait for the ability to decrypt it. The time to protect long-lived secrets is before they are intercepted.
Oracle AI Database 26ai introduced quantum-resistant cryptography a year ago, with capabilities that include post-quantum and hybrid key exchange for TLS. More recently, we brought TLS 1.3, post-quantum cryptography, and FIPS 140-3 mode to Oracle Database 19c, starting with Release Update 19.32. The new cryptographic assessments help customers identify the upgrades, configuration changes, and client-compatibility testing needed to adopt these capabilities.
Security at the source gives every application a stronger foundation—including the applications you are yet to build.
Secure through Resilience
Ransomware, operational mistakes, and infrastructure failures can interrupt even well-protected systems. A backup is useful only if you can trust it, restore it, and resume the business.
Oracle Maximum Availability Architecture combines reference architectures and operational practices with technologies such as RAC, Active Data Guard, and GoldenGate. Customers can design for planned maintenance and unexpected outages according to their availability and recovery requirements.
Oracle Zero Data Loss Recovery solutions provide capabilities designed to support real-time transaction protection, incremental forever backups, and immutable backups across supported cloud and on-premises environments.
We are announcing backup replication across selected OCI regions to support recovery after a primary-region outage. To help teams investigate recovery-service operations, a new MCP server lets AI clients generate dashboards and reports on capacity, performance, and incidents. Add link to the blog.
Put the strategy to work
Oracle’s security strategy extends to fleet-wide assessment, policy management, patching, and recovery. KuppingerCole recognized Oracle as an Overall Leader in its 2025 Leadership Compass for Data Security Platforms. Steve McDowell, Chief Analyst and CEO of NAND Research, described our framework as “the most comprehensive security architecture currently being brought to market.”
Start putting that strategy to work today:
- Apply the October release updates. A security fix cannot protect your database until it is deployed. Promptly apply Database Release Update 23.26.4 to Oracle AI Database 26ai and 19.33 to Oracle Database 19c, before adversaries reverse-engineer fixes and exploit unpatched systems. Include clients, infrastructure, and non-production environments. To reduce the ongoing burden of keeping databases current, consider Oracle Autonomous AI Database, where Oracle manages database patching for you. Consult the October Critical Patch Update advisory.
- Use the tools available to you. Our free and heavily discounted security and patching tools help reduce procurement barriers to assessment, patch automation, testing, and validated switchovers. Free offers run through February 28, 2027; qualifying one-year patching and upgrade licenses are discounted by 90%, and the offer is valid through May 31, 2027. Applicable support requirements and offer terms apply.
- Verify protection across the fleet. Assess exposure and address the identified risks, check application and agent authorization, and test recovery.
Every database should be part of your organization’s security plan. Use these capabilities to close today’s gaps and extend protection to the applications and agents you deploy next.
Related blogs:
Oracle Deep Data Security Expands Authorization Across Analytics, APEX, ORDS, and AI Agents
Make cryptographic posture visible across your Oracle Database fleet
Introducing Data Subsetting in Oracle Data Safe for Smaller Datasets
