VMware Tanzu delivers speedy, AI-ready data access for AI agents

Broadcom Inc. is expanding the capabilities of its VMware Tanzu Platform for developers, adding support for artificial intelligence agents via a new data lakehouse called VMware Tanzu Data Intelligence, which provides low-latency access to multimodal data at large scale.

It was introduced today at VMware Explore 2025 in Las Vegas alongside VMware Tanzu Platform 10.3, a new update that promises to help integrate enhanced generative AI capabilities into new and existing applications, Broadcom said.

Officially debuted in 2024, Tanzu Platform is a Cloud Foundry-based platform as a service that provides a pre-engineered AI application development platform with built-in best practices, configurations and optimizations that allow developers to focus on building and deploying applications.

Delivering AI-ready data

With the release of VMware Tanzu Data Intelligence, Broadcom is now turning its attention to the development of agentic AI applications that can automate complex business processes and tasks with limited human intervention. Broadcom said it’s designed to integrate with Tanzu Platform in order to provide AI with simple, low-latency access to real-time data, no matter where it’s generated. In other words, it’s meant to feed AI agents with the essential real-world information they need to be able to operate effectively.

Earlier today, the company announced a revamped VMware Cloud Foundation platform for developing AI-native private clouds, and Tanzu Data Intelligence builds on that release. According to Broadcom, it’s basically a modern data lakehouse platform, with comprehensive capabilities for data ingestion, processing, querying and AI enablement.

The lakehouse can handle diverse data types from any source, including structured and unstructured, and native or federated information, with the ability to scale to petabytes in volume with latency measured in milliseconds. It supports “massive concurrency” across users, data and application programming interfaces, with full data lineage for sovereignty and governance. What this means is that users can easily understand what data informed an AI outcome or decision, enhancing observability and oversight.

According to Broadcom, VMware Tanzu Data Intelligence supports almost every kind of AI application, including transactional, smart and agentic workloads, model training, fine-tuning and decision support. It also boasts native vector search too, enabling querying and semantic similarity search across vectorized unstructured data.

In addition, VMware Tanzu Data Intelligence comes with a number of novel features aimed at AI application development, including ingestion and workflow orchestration capabilities that support data movement and transformation from any source to create real-time data flows and event-driven architectures. Its federated query services ensure simple access to varied data sources, and it can access Tanzu Platform’s container compute services to scale its compute needs elastically as required.

Other components include in-memory and transactional data layers, enabling real-time data services for operational intelligence, real-time decision-making and live digital experiences, plus advanced analytics tools such as predictive models, vector processing and more.

Purnima Padmanabhan, vice president and general manager of the Tanzu Division at Broadcom, said many organizations continue to struggle with fragmented and siloed data and the high and unpredictable costs of moving it between different environments. This is exactly what the company is trying to solve.

“Tanzu Data Intelligence combined with Tanzu Platform provides our customers with the power of a modern multimodal data lakehouse architecture and our proven single-command CF-push technology but with the cost predictability of a private cloud,” she said.

Tanzu Platform updates

As for the Tanzu Platform 10.3 release, it’s being billed as a “prescriptive, pre-engineered AI application development platform” with integrated best practices, configurations and optimizations. By abstracting away headaches such as provisioning servers and networking monitoring, it frees developers to focus on building their AI apps.

The new features include granular AI model service plans, complete with new quota capabilities and broader integration via webhooks, so users can better control security and cost management. The release also adds service publishing in the Tanzu Platform Marketplace, giving developers a way to sell their AI applications as services, while the new application onboarding and modernization tools will help users to identify, refactor and automate app migration and modernization with improved performance and lower costs.

Elsewhere, there’s a new vulnerability dashboard in the Tanzu Platform for visualizing risk exposure for Tanzu-based applications and platform components, and integrated data services observability for Tanzu Hub, providing insights into performance, trends and bottlenecks, the company said.

Finally, to simplify AI application development further, VMware Tanzu announced a new “AI Starter Kit” for developers, complete with tooling and custom code for automating the standup of Tanzu AI services and how-to guides for application deployment.

Images: VMware