The hidden risk of AI ambition: Why your infrastructure is holding you back
By Daryush Ashjari, Chief Technology Officer and Vice President of Solutions Engineering, Nutanix Asia Pacific & Japan
Over the past decade, the view from the boardroom has shifted at breakneck speed. Ten years ago, the “C-suite mandate” was cloud computing. Five years ago, it was the wholesale shift to SaaS (Software as a Service). Today, AI isn’t just a mandate – it’s the gravity pulling every business decision toward it.
CEOs and boards are hungry for the “AI edge”, pushing for radical innovation and hyper productivity. But there is a silent friction developing beneath the surface: AI ambition is accelerating faster than the digital foundations built to support it. So what got us here won’t get us there!
According to Nutanix’s latest Enterprise Cloud Index (ECI) Report, between 77% and 87% of enterprises across the Asia-Pacific region are struggling with infrastructure that’s not ready for AI workloads.
For the C-suite, this isn’t just a technical hurdle; it’s a strategic bottleneck. The ability to operationalize AI at scale increasingly depends on overcoming constraints around cost, data sovereignty, hardware availability, and software complexity. These underlying factors will often determine whether AI ambition becomes a competitive advantage or a lurking beast in the shadows.
AI Is Pushing Infrastructure Beyond the Data Center
For decades, enterprise computing was predictable: applications lived in centralized data centers or public clouds, managed by dedicated IT teams. AI has shattered that model.
According to the Enterprise Cloud Index, more than four in five executives say their infrastructure is not ready for GenAI. And the number one reason it stalls before it scales is integration. Most environments were built for predictable workloads. AI is the opposite; it is continuous, data-sensitive, resource-hungry, and unforgiving.
We are seeing a massive shift toward “distributed intelligence.” Workloads now need to live at the edge, closer to where the pulse of the business actually beats. Think of retailers running real-time analytics on the shop floor, or mining companies deploying AI-powered safety systems at remote sites with zero on-site IT staff.
This new reality demands an environment that functions consistently across data centers, public clouds, and the edge, without forcing teams to reinvent the wheel for every new location.
Containers have become the “universal language” of this flexibility. By packaging applications, AI models, and dependencies into portable units, containers allow workloads to move seamlessly across environments. It is therefore not surprising that 85% of global enterprises say AI initiatives are accelerating container adoption. In India, that sentiment is near-universal at 97%. But here is the catch: traditional infrastructure was never designed to manage this level of fragmentation.
Shadow AI Is a Symptom of a Deeper Infrastructure Gap
When IT can’t move at the speed of business, the business moves without IT. Shadow IT was a proving point, and now we face a new phase – Shadow AI.
Today, deploying non-IT-approved AI capabilities can be as simple as swiping a credit card. Our ECI report reveals that nearly four in five IT leaders across APJ report encountering AI solutions deployed outside formal IT oversight. This phenomenon, often referred to as shadow AI, is frequently framed as a governance issue. In reality, it is also an infrastructure problem.
Every unsanctioned deployment signals a mismatch between what the business wants to build and what IT environments can currently support. In Singapore, for instance, 93% of organisations say silos between business units and IT make effective technology execution difficult. When companies focus only on restricting shadow AI rather than addressing this underlying gap, they end up managing symptoms instead of solving the root cause.
The pattern is not new. During the early days of cloud adoption, shadow IT emerged in much the same way. Enterprises that responded with tighter controls often slowed innovation without resolving the structural problem.
With AI, however, the stakes are significantly higher. Imagine if your business units develop 30 AI use cases and applications on their own with no governance. What would be the impact on your IT operations team? On the other hand, regulatory frameworks such as Singapore’s MAS guidelines,
Australia’s Privacy Act reforms, and India’s expanding data protection rules are raising expectations around transparency, governance, and accountability in AI systems. Shadow AI, by its nature, makes meeting those expectations far more difficult.
A sobering fact: 20 AI-based use cases translate to at least 117 open source components. Each of these open source components has its own release cycles, patch updates, and so on. We can clearly see the impact this has on IT. It will become a wild, wild west very quickly.
AI Experimentation Is Forcing a Rethink of Enterprise IT
And yet, that’s not stopping anyone from dreaming big. Enterprises are rapidly experimenting with agentic AI to automate routine processes, streamline workflows, and unlock new sources of productivity. From customer support to internal operations, organizations are exploring how AI agents can handle increasingly complex tasks with minimal human intervention.
But that’s not all. Our research shows that 58% of IT leaders globally expect AI agents to improve productivity, and an almost equal 57% believe they will enable new products and revenue streams. Across APJ, leaders increasingly view agentic AI not simply as an efficiency tool, but as a catalyst for new commercial opportunities.
Scaling these experiments requires a redesign of IT. Companies navigating this shift are moving from a gatekeeping model toward building platforms that enable innovation while embedding oversight.
The framework for the future-ready enterprise rests on two pillars:
Proactive IT enablement: Don’t wait for the business to ask. Develop a platform, built on a secure, governed “innovation playground” from the start. If the official platform is faster and easier than the “shadow” version, the business will use it.
Built-in governance and monitoring: Embed auditing, compliance, and policy enforcement directly into the platform. This allows IT leaders to detect risks early, engage constructively with business teams, and maintain security without discouraging experimentation.
Supporting this approach requires a unified infrastructure designed for hybrid environments. Platforms must operate consistently across data centres, public clouds, and edge locations while supporting both traditional virtual machines and modern containerized workloads. When implemented effectively, this model allows companies to experiment with AI at speed while maintaining enterprise-grade control.
Future-Proofing the AI-First Enterprise
AI adoption is no longer a race to see who can build the most models; it is a race to see who can actually operationalize them.
As organisations move from “sandbox” experimentation to deployment, the underlying infrastructure will be the ultimate arbiter of success. In this new AI era, competitive advantage will not come from adopting AI first. It will come from building the technology foundations that allow AI to scale without breaking.