Why control is an asset in the AI era

Before you read another word, ask yourself four questions:

1. Do you know exactly where your AI models were trained — and whose infrastructure they ran on?
2. Could you move a critical workload somewhere else in a few weeks without creating a compliance headache?
3. Who can access your most sensitive systems, and under whose jurisdiction do they operate?
4. Can you explain how your models are governed, from the data that trains them to the outputs they produce?

If any of these questions made you hesitate, you may have less control over your AI environment than you think.

The cloud era was characterized by removing constraints so you could move data wherever it needed to go, scale applications across regions and remove friction. The goal was “build once, run anywhere.”

Now, more demanding regulations and geopolitical uncertainty are making the location, ownership and control of technology harder to ignore.

Why organizations are designing for control

Control is something you can lose gradually — one dependency at a time — until a choice you thought was yours no longer is. And that risk increases every time you move a workload, give a model access to sensitive data or become more dependent on a single provider.

When you keep your intellectual property (IP) within boundaries you choose and maintain the freedom to move workloads, models and data as circumstances change, control becomes a business advantage.

Organizations are therefore deliberately designing for control. Our global cloud research, Cloud-led innovation in the era of AI: The new rules for driving value with cloud, shows that nearly all organizations (99%) expect private cloud adoption to increase. When determining where to place AI workloads, 51% point to data security, privacy and compliance as leading considerations, while 49% cite the need for greater control over infrastructure and customization.

Sovereignty has switched from compliance to strategy

Sovereignty used to be something governments worried about, or something regulated industries left to their legal and compliance teams. Now it has become a board-level strategic priority, with the geopolitical climate adding to the urgency. Our 2026 Global AI Report: A Playbook for Private and Sovereign AI notes that 59% of AI leaders — organizations seeing the strongest AI-linked business performance — rank sovereignty as their top governance concern.

This makes sovereignty a question of resilience as much as compliance. You need to understand where your critical dependencies lie, what risks those dependencies create and how exposed they would be if the conditions around them changed.

Why AI requires a different level of control

It’s one thing to run an application in someone else’s environment. It’s another to feed an AI model your customer data, proprietary knowledge, IP and years of accumulated expertise — particularly when you don’t have complete control over where that information goes or how it is handled.

This helps explain why 95% of organizations in our AI report see private or sovereign AI as important to their AI strategy.

But wanting greater control and achieving it are two different things. About 35% of CAIOs identify the difficulty of building, integrating and managing complex AI models in private and sovereign environments as their top barrier to AI adoption.

Just deciding that sensitive AI needs greater protection isn’t enough. You also need to create an environment where you can actually build and scale it.

Sovereignty is not just about data

So, what does sovereignty mean in this environment?

Conversations about sovereignty used to start and end with data residency. Now, you also need to know who controls the infrastructure beneath your data, who operates the services you depend on, which technologies you’re tied to and how your AI is governed.

In practice, control now extends across four dimensions of sovereignty:

1. Infrastructure sovereignty: Control over the computing, network and platform environments on which your organization depends.
2. Data sovereignty: Control over where your data resides, how it moves and how it’s processed.
3. Operational sovereignty: Control over how critical digital services are operated, accessed and governed.
4. AI sovereignty: Control over how AI models, data, algorithms and intelligence are developed, deployed and governed.

How much control have you retained over the technology that turns your data into intelligence?

Hybrid cloud is the practical answer

The next question is where all of this should run. There isn’t one answer, because not every workload needs the same level or type of control.

A highly sensitive AI model trained on proprietary data may belong in an on-premises or private environment. A regulated workload may require a sovereign cloud with specific jurisdictional and operational assurances, while other applications may benefit most from the scale, broad ecosystems and rapid access to new capabilities available in public cloud.

Hybrid cloud offers a solution: Rather than forcing everything into the same destination, you can use the environment that best suits the requirements of an individual workload — balancing control, security, sovereignty, performance, cost and access to innovation according to what that workload needs.

Modernization is a control strategy

Once you’ve made those placement decisions, you need to make sure your environment can keep adapting as regulations and business requirements change, new AI capabilities emerge and providers change their services and commercial models.

Modernization is how you build that adaptability into your technology foundation. You need infrastructure that can scale AI securely and efficiently, and platforms that provide:

• Flexibility: You can move workloads as business requirements change.
• Governance: Critical data and AI remain within defined boundaries.
• Resilience: Avoid becoming overly dependent on any single environment or provider.
• Observability and automation: Manage complex hybrid environments.

Risks that once seemed remote can become immediate, which is why modernization must be an ongoing discipline in your organization.

Manufacturer accelerates AI development by first establishing control

One of our clients, a European manufacturer operating in a highly security-sensitive industry, wanted to give teams from multiple business entities the computing power to train, fine-tune, test and operationalize AI models at scale — all under a single governance model, and without their sensitive data leaving their own environment.

Working with leading technology partners, we designed and deployed an on-premises AI platform that allows more than 100 data scientists, from multiple entities, to access a shared, secure environment.

Designed for data confidentiality and high availability, the platform — which we continue to run and optimize under a five-year agreement — launched with a 32-GPU cluster. It provides a scalable foundation for extending AI throughout the organization and enables faster development of AI and large language models, with shorter training cycles.

The manufacturer didn’t trade control for speed. By creating an environment where their data could remain protected while teams had the infrastructure they needed to work at scale, greater control became an enabler of speed.

How we’re helping you navigate the need for control

Go back to the four questions at the beginning. Where was your answer a confident “yes,” and where was it closer to “I’m not sure”? Your answers will reveal where decisions about your infrastructure, data, operations or AI may no longer be entirely yours to make.

However, there isn’t one boundary that makes sense for everything. NTT DATA creates AI-ready environments designed around the level of control each workload needs. We bring together hybrid cloud modernization, sovereign cloud, governance, security and managed services so you have choices when your needs, your industry’s regulations or the world around you change.

Sovereignty and innovation aren’t opposing forces. With the right architecture and infrastructure, greater control gives you the confidence to use AI in places where the risks might otherwise be too high.

What to do next

Ready to build an AI foundation you control? Talk to us about our Hybrid Cloud Modernization solutions.