Delivering meaningful business value from enterprise AI relies on the AI’s secure access to the enterprise’s proprietary data. Unfortunately, many organizations building out AI initiatives come to find that data is often locked away in disjointed systems. For data to be truly ready for AI, it can’t just sit in a silo. Useful data needs to be fresh, secure, and instantly available so that AI agents and people alike can make better decisions faster. Messy data and broken pipelines create more than IT headaches; now they are also creating roadblocks that prevent teams from making AI work. Tanzu Data Intelligence was built to address this exact challenge for the private cloud.
In a previous VMware Tanzu Data Intelligence update, we focused on foundational unification, bringing together data at rest and data in motion to support next-gen applications. We introduced vector capabilities for AI use cases and streamlined the integration between operational and analytical data layers.
Today, we’re releasing Tanzu Data Intelligence 10.4, which features several new capabilities that can help simplify data access, reduce manual administration, improve platform resilience and optimize your AI investments. These updates are designed to improve how teams interact with, manage, and secure their data environments.
Here is an in-depth look at the benefits that can be derived from the latest updates to Tanzu Data Intelligence.
Unlock insights from data faster with AI-powered natural language search and data access
Data analysts typically spend hours crafting complex SQL queries to extract insights. This release introduces the SQL Assistant for Tanzu Data Intelligence, a new capability that empowers users to query data using natural language prompts instead of manually writing SQL code.
The SQL Assistant allows users to interact with their data using natural language questions such as, “Show sales trends by region for Q3 compared to last year.” The AI-powered SQL Assistant system automatically translates the natural language search query into a SQL query for VMware Tanzu Greenplum and fetches the answer.
Analysts may choose to copy the AI-generated SQL query into an editor for refinement, or they can bypass SQL entirely and rely on the natural language answers. This can improve data analysts’ speed and efficiency, help junior analysts understand legacy code, provide optimization suggestions for massively parallel processing (MPP) workloads, and reduce the time-to-insight for business users with little to no SQL experience.
Automate data administration with a hardened MCP server
Managing a data warehouse requires specialized expertise. To streamline these complex operations, we are introducing a new VMware Tanzu Greenplum MCP server that wraps multiple critical system insight functions into a secure, AI-friendly API. This new feature empowers AI-assisted data platform administration, reducing the burden on internal resources.
Database administrators can now use natural language to interact with the platform. Whether it is querying system health, discovering underutilized data that can be safely cleaned up, or identifying the exact system parameters to tune for optimal performance, the MCP server serves as a single, unified point for automated platform management and health checks.
Read more about how the Tanzu Greenplum MCP server puts security patterns into practice here.
Achieve data sovereignty, resilience, and cost savings at scale on private infrastructure
Deploying autonomous AI agents introduces risks that public cloud services cannot mitigate. Bring AI compute and data together within a more secure, private environment with cloud-like agility and cost benefits.
Deploy sovereign data architecture faster with an enhanced VCF installer
Deploying a multi-node data warehouse on premises typically involves extensive planning, manual configuration, and coordination across infrastructure and platform layers. The new VMware Cloud Foundation installer for Tanzu Data Intelligence streamlines this process, reducing deployment time from weeks to hours. The installer provides a guided, GUI-driven workflow that captures key configuration inputs and automates the end-to-end setup, including infrastructure integration and recommended system configurations. As a result, clusters are provisioned with consistent, production-ready defaults, minimizing manual effort and reducing the risk of misconfiguration.
Eliminate downtime with frictionless “in-place” upgrades
Migrating major data warehouse software versions (e.g., from Tanzu Greenplum 6 to 7) has historically been complex, involving massive data export and import—a process that can take days for petabyte-scale clusters. The new, in-place upgrade utility streamlines the experience of upgrading the cluster without moving the data. It reduces the downtime and disk overhead of data movement. That means you can get to the latest features of Greenplum 7 faster, with the ability to roll back if needed before the final commit.
Reduce cost and limit “blast radius” during disaster recovery with selective object filtering
While disaster recovery (DR) of enterprise data is absolutely critical, replicating petabytes of data for DR is expensive, bandwidth intensive, and time consuming. Yet in many cases, only some data needs to be replicated for continuity in the event of a disaster, while other data sets can often be reproduced from source, with no replication required.
Now teams can better optimize data replication for DR in Tanzu Data Intelligence. The newly enhanced Tanzu Greenplum Disaster Recovery (GPDR) supports selective object filtering, enabling administrators to filter replication at logical boundaries, such as excluding specific databases, schemas, or tables. This new enhancement helps with optimization in environments where full replication is impractical by reducing storage costs and network bandwidth usage, only replicating required data to the DR site.
Enable built-in compliance for data at rest with Transparent Data Encryption (TDE)
Securing mission-critical data requires anticipating worst-case scenarios, including physical theft or unauthorized operating system access. In this regard, we are introducing native Transparent Data Encryption (TDE) to secure your data at rest. Administrators can seamlessly enable TDE during the initialization of a new cluster and use their existing Key Management Service (KMS). Once initialized, every piece of user data is automatically encrypted. Even if a bad actor manages to bypass physical security and copy the raw data directory directly out of the cluster, the data would be fully encrypted on the disk.
Apache Iceberg support (technical preview)
Data teams often grapple with the sheer cost and complexity of duplicating vast amounts of cloud data into a central warehouse just to run analytics. We address this “data gravity” challenge by bringing the database to the data. With a new feature available as a technical preview, Tanzu Greenplum can directly query Apache Iceberg data stored in S3. Instead of building complex ETL pipelines to move information into a central warehouse, Tanzu Data Intelligence empowers teams to analyze data exactly where it sits. By tapping into this open table format, you can reduce expensive data movement, accelerate the launch of new analytical projects, and make data sharing between Tanzu Data Intelligence and your other systems more seamless, more secure, and significantly faster.
Realize more value from your data and AI with agent-ready data streaming and synchronization
AI agents need enriched context, working memory, and coordination of multi-agent swarms to avoid conflicts.
Synchronize multiple apps or agents with a new distributed types extension
As applications and agents scale across distributed environments, keeping their shared state synchronized becomes a major challenge. When multiple systems try to read, update, or coordinate around the same data simultaneously, it often creates data inconsistencies and forces teams to stitch together complex, external coordination tools.
VMware Tanzu GemFire now includes a supported extension for distributed data structures, enabling common coordination patterns to be implemented directly within the data grid.
This extension introduces primitives, such as distributed lists, sets, AtomicLongs, and semaphores, all built on top of Tanzu GemFire’s existing partitioning, replication, and fault-tolerance mechanisms. These structures integrate with the platform’s consistency and failover model, providing predictable behavior across nodes in a distributed system.
For developers, this makes building distributed systems much easier. Instead of juggling external tools or writing custom code to keep multiple apps in sync, you can manage your data and coordination logic all in one place.
This approach is particularly useful for Java-based systems that require distributed coordination, such as workload orchestration, rate limiting, resource management, and stateful microservices, while maintaining alignment with enterprise-level operational requirements.
Empower agents and apps by bridging real-time, event driven operations with large-scale analytics
Operational messaging systems and big data analytics traditionally live in silos, separated by high-latency pipelines that make closed-loop, real-time reactions (for example, stopping fraudulent transactions mid-flight) extremely difficult.
To better integrate real-time event processing with large-scale analytics, VMware Tanzu RabbitMQ introduces enhanced integration patterns with Apache Spark and streaming data pipelines.
A new bidirectional integration with Apache Spark enables event streams from RabbitMQ to be consumed directly into Spark-based processing jobs for enrichment, aggregation, and anomaly detection. Processed results can then be published back into RabbitMQ, allowing downstream services to react in near real time. This supports closed-loop workflows, such as fraud detection, where events are analyzed at scale and decisions are fed back into operational systems with low latency.
On the developer side, the introduction of the RabbitMQ Stream Browser provides a built-in way to inspect and debug stream data directly at the broker level. This reduces reliance on custom consumers or external tooling when troubleshooting message flow, offsets, and payloads.
Additionally, enhanced JMS Message Selector support allows Java applications using JMS APIs to perform server-side filtering of messages based on headers and properties. This enables more efficient message routing and helps existing JMS-based systems integrate with RabbitMQ without requiring significant changes to application logic.
Together, these capabilities provide a more cohesive model for combining event-driven systems with batch and streaming analytics, while maintaining compatibility with existing enterprise messaging patterns.
Bringing it all together: A simpler, more connected data platform
Tanzu Data Intelligence enables data access, processing, and coordination to happen entirely within your secure boundaries, giving you greater control and privacy over your most sensitive assets. AI-assisted query and administration features reduce the overhead of working with large schemas and complex systems, helping teams move from intent to execution more efficiently, while improvements in deployment, upgrades, and encryption address the operational realities of running data platforms on private infrastructure with a focus on reliability, repeatability, and control. Ultimately, Tanzu Data Intelligence empowers your teams to stop moving data and start acting on it, unlocking valuable insights faster and building the next generation of intelligent applications with absolute confidence.
