Establishing a Multicloud Data Strategy for the AI Era

In my experience working with enterprise leaders, the journey to the cloud rarely follows a straight line. Many organizations set ambitious goals to move all operations to the cloud. They quickly find that certain legacy systems must remain on-premises.

This reality results in a complex, hybrid multicloud environment. That means they need to adopt a new strategy for managing test data. Enterprises have to navigate the challenges of integrating old and new, balancing innovation with the demands of legacy infrastructure.

You might already split modern workloads on AWS or Azure using their compute and platform services. At the same time, you may need to keep core legacy databases firmly on-premises. To make matters worse, there might be a different combination across your many geographies. 

This kind of fragmented landscape presents significant bottlenecks and complex administrative overhead.

To accelerate application delivery and pursue digital transformation, you need fast, trusted, AI-ready data — accessible across all your environments.

The Realities of Modern Cloud Architectures

The shift to multicloud architectures brings immense flexibility. However, it also introduces deep complexities for your data orchestration teams.

Why Do Hybrid States Persist in Cloud Strategies?

I regularly speak with organizations attempting to fully abandon their on-premises infrastructure. Inevitably, they encounter significant obstacles. Certain legacy applications have intricate dependencies and are either too complex or too costly to migrate.

As a result, organizations have to support both traditional on-premises databases and modern cloud deployments at the same time. In this landscape, data silos are common and hinder agility.

Why Are Non-Production Environments Left Behind in Cloud Transformation?

In my experience, as production applications advance in the cloud, non-production environments do not necessarily reap the same benefits. Too often, environments are shared across many dev teams to save costs, static long-lived test environments are forgotten, and incomplete testing occurs due to limited orchestration with legacy databases. 

Transforming these non-production environments is essential. By focusing on modernization here, you unlock the true value of both hybrid and multicloud strategies. Agile, automated, and up-to-date test environments are a critical factor in meeting business demands.

How to Handle Heterogeneous Sources 

Operating in a multicloud environment means dealing with diverse data sources. For example, you may have an on-premises Oracle database, a PostgreSQL instance running on AWS, and a Snowflake environment used for analytics.

Each platform comes with its own unique interface, security protocols, and methods of connection. These differences can become major hurdles for IT and development teams.

Modern automation can help overcome these obstacles:

• Kubernetes-native automation: Spin up containerized databases in seconds.
• Rapid environment provisioning: Quickly provision or clone test environments without manual intervention.
• Improved efficiency: Reduce the time your developers spend managing infrastructure, so they focus more on writing code.
• Isolated and coordinated deployments: Provision safe copies of these data sets in isolation or as a coordinated group.

Without these advances, teams waste valuable time wrestling with infrastructure chaos. Embracing automation strategies lets you maintain agility in a heterogeneous cloud landscape.

Want to Become a Modern, Cloud-Driven Enterprise?

If you're balancing on-premises and cloud data, watch this recent webinar from myself and Daniel Stolf. You will learn how to best optimize your test data management between environments and save on storage costs.

How to Address New Threats to Your Multicloud Estate in the AI Era

AI transforms how we interact with information. While this brings incredible speed to development, it also introduces serious data privacy and compliance risks.

What Risk Does Agentic AI Pose?

Agentic AI tools scan massive swaths of data at speeds no human can match. When you connect an AI agent to your corporate environment, it identifies information that human workers may have forgotten even exists — or unknown security holes.

This speed and thoroughness introduce a major new attack vector. If sensitive production data leaks into non-production environments, an AI agent could surface private information during a simple, routine query. 

The risk of unintentional exposure increases drastically as AI-driven automation becomes standard and you further split your data across on-premises and cloud.

How to Secure Sensitive Information?

To innovate with confidence, you must eliminate sensitive data risk across all multicloud environments. Delphix automatically discovers and protects sensitive data through automated data masking and policy-based governance. 

By centralizing control, you reduce regulatory risks and ensure compliance with strict privacy laws like HIPAA and GDPR. Downstream teams can access fresh, high-quality masked data on-demand.

Building a Unified Data Fabric: Where an Enterprise Test Data Management Platform Becomes Critical

Enterprises cannot solve modern data challenges with a patchwork of point tools. You need a single, secure platform that unifies data delivery, governance, and compliance.

How Can You Orchestrate Data Across Environments?

As applications span on‑premises systems, public cloud, and managed services, data orchestration directly affects delivery speed. Teams need reliable access to data at every stage of development. Without a coordinated approach, fragmented data, slow refresh cycles, and inconsistent environments quickly become a bottleneck.

To manage this complexity at scale, enterprises need a centralized way to coordinate data delivery, consistency, and control across all environments.

Perforce Delphix is the intelligent data automation platform built for enterprise scale and complexity. Delphix creates a unified data fabric that seamlessly spans on-premises systems, cloud infrastructure, and PaaS environments. 

Whether your data lives in a legacy Oracle database or a cloud-native Databricks environment, Delphix gives you a single point of control. You can refresh, rewind, bookmark, and branch data through a variety of simple  interfaces, such as an MCP Server.

Image

How to Supply AI-Ready Data?

To speed up scenario testing and unblock new development, your teams need highly accurate test data. That data must be realistic, safe to use, and available on-demand as AI-driven workflows place more pressure on non‑production environments. 

With Delphix, teams combine automated masking with secure synthetic data generation. They can deliver production‑like synthetic data alongside masked copies to cover rare edge cases while preserving referential integrity across databases and warehouses.

Image

Get the Report

Your Peers Have Spoken.

Perforce is a Customers’ Choice in the Gartner® Peer Insights™ 2025 Voice of the Customer Report for Test Data Management (TDM).*

Access the Report

_{* Gartner, Gartner Peer Insights ‘Voice of the Customer’:  
Test Data Management, Peer Contributors, August 2025} 

Gartner, Gartner Peer Insights Voice of the Customer, Test Data Management, Peer Contributors, 30 July 2025,

Gartner, Peer Insights and the Gartner Peer Insights Customers' Choice badge are trademarks of Gartner, Inc. and/or its affiliates.

Gartner Peer Insights content consists of the opinions of individual end users based on their own experiences, and should not be construed as statements of fact, nor do they represent the views of Gartner or its affiliates. Gartner does not endorse any vendor, product or service depicted in this content nor makes any warranties, expressed or implied, with respect to this content, about its accuracy or completeness, including any warranties of merchantability or fitness for a particular purpose.

*IDC Business Value White Paper, sponsored by Delphix, by Perforce, The Business Value of Delphix, #US52560824, December 2024 

This blog was originally published by Matt Yeh and has been updated by Nick Mathison.