In the quest to deploy artificial intelligence at scale, enterprises are discovering a basic truth: Model capability is no longer the only issue, but data access and control can determine success or failure for AI initiatives.
This realization has led to a focus on AI infrastructure, the buildout of a data foundation that can provide unified access, metadata intelligence, security and hybrid orchestration. The next phase of AI-driven transformation will rely heavily on how organizations use data platforms to turn AI proofs-of-concept into production value.
AI stacks are rewriting the rules of business, transforming core mechanics of the enterprise. Infrastructure matters when deploying AI models, determining how decisions get made, how work gets executed and how risk gets governed.
“The winners will be the organizations that build a complete system around those models – one that connects to existing deterministic applications, creates a shared truth layer, controls agents as they take action, and uses human feedback to continuously improve,” said theCUBE Research analysts Dave Vellante and George Gilbert. “This is why we believe the upside is so large. Enterprises that get this right won’t just run cheaper — they will run differently. They will scale with less proportional labor growth, compress cycle times from insight to action and start to behave more like platform companies, with compounding advantage that is difficult for competitors to copy.”
This feature is part of SiliconANGLE Media’s exploration of the architectural shifts powering continuous, production-grade AI. Be sure to check out theCUBE’s “Dell AI Data Platform Event: From Ambition to AI at Scale.”
Modular solutions for data access
The process of moving AI from pilots into production requires a stronger, yet adaptable data foundation. Larger enterprises are using different infrastructure and deployment models for a wide range of workloads, and this is shaping how key compute providers such as Dell Technologies Inc. are delivering infrastructure solutions.
Composability has become more popular. Rather than adopting a single stack, enterprises are seeking architectures that can support plug-and-play data engines and frameworks.
“What we’re finding is customers want modular solutions,” said Varun Chhabra, senior vice president of product marketing and infrastructure solutions group at Dell, during a recent interview with theCUBE, SiliconANGLE Media’s livestreaming studio. “They want to think about [AI] across the whole platform. Compute, storage, networking, GPU. The software framework on top of it, the models, have they all been tested, have they all been validated?”
Model validation and modular solutions are becoming even more essential with the rise of agentic AI. The goal for many enterprises today is to connect information with desired behaviors for agents and other AI tools. This requires an architecture that can feed agents valuable proprietary data for measurable results. Dell and other firms are building infrastructure for agents in the data layer to help companies realize this vision.
“The agent itself is just a software system, and it has a number of components … it has LLMs, it has knowledge graphs, it has protocols,” said John Roese, global chief technology officer and chief AI officer of Dell, during an interview with theCUBE. “That data layer is not magic. It’s a real thing. You have to actually build it, you have to build an infrastructure that supports knowledge graphs and maintains them and can feed them, that also can do things like agentic.”
Focus on storage and security
AI’s expanding influence in the enterprise stack and the demands of scalable data management are also shifting requirements for storage, security, sovereignty and governance. The transformation of storage infrastructure has been significant. In May, Dell announced enhancements for its PowerStore platform that Chhabra described as “the biggest leap forward in the platform’s history.”
The changes included hardware and software enhancements that could deliver up to three times more input/output operations, throughput and density than previous generations.
“It is a new class of modern data platform built to help customers lead through change, not just react to it,” Chhabra said.
Security has also become a front-burner issue in the age of AI. The development of new context-aware, autonomous protection tools has highlighted for many security practitioners the potential for securing the enterprise data platform in more advanced ways.
“Understanding an event may require security context, production behavior, application dependencies, and business impact to come together at the point of decision,” said theCUBE Research analyst Krista Case. “That makes context part of the security architecture rather than an enrichment step added after an alert fires, and it changes how enterprises should evaluate AI security capabilities. It makes a critical difference if the system has comprehensive visibility across the organization’s environment, understands where that knowledge comes from, and is current enough to support the decision being made.”
There are also concerns that AI has opened a large security hole for organizations. An independent research study from Omdia commissioned by Dell found that 79% of organizations have already experienced an AI-related incident in the past 12 months. Security analysts have documented numerous instances in which AI can accelerate attacks and expand the attack surface around enterprise systems and data.
“Economics and the operating models surrounding security operations are changing,” Case said. “Attackers have new ways to reduce the time and expertise required for portions of their work, and defenders have an opportunity to remove human coordination from portions of theirs.”
Exercising provable control
Defending the data layer now goes beyond monitoring networks for potentially malicious behavior. Supporting AI at scale requires the ability to operate across multiple platforms and draw on different data sources to ensure accountability.
This marks a shift in the enterprise, with AI governance is moving from observability to provable control, according to theCUBE Research’s Paul Nashawaty. As agentic workloads move across infrastructure, organizations increasingly need to prove what the agent was authorized to do, why it was allowed to take a specific action and whether that authority remained intact.
“Enterprise AI governance is moving from a visibility problem to an accountability problem,” Nashawaty said. “The objective should not be finding a single AI governance product that claims to solve every layer of the problem; it should be creating an architecture in which authority, policy, enforcement, visibility and evidence remain connected.”
The need for stronger governance has also given rise to sovereign AI, the strategic capacity of a government or private enterprise to independently manage, develop and operate its AI lifecycle with full authority over data, compute, models and policy.
Organizations are increasingly turning toward sovereign solutions to deploy AI responsibly and at scale. A global IDC study commissioned by Dell, found that 52% of government respondents plan to invest in sovereign AI within 12–18 months, and 58% identify strong sovereign data governance, quality and control as among the most critical platform requirements for sovereign AI.
“Sovereignty adds another layer,” Nashawaty said. “For organizations operating in regulated, disconnected or air-gapped environments, governance capabilities may need to run entirely within customer-controlled infrastructure.”
Implementing the AI Data Platform
Control of infrastructure revolves around data, and key enterprise tech companies such as Dell are providing a foundation for the era of intelligence through offerings such as the AI Data Platform.
Dell’s platform is designed to let enterprises access data across silos and move swiftly from proof of concept to production by using several key architectural pillars. These include storage engines such as PowerScale and ObjectScale for unstructured and semi-structured data, solutions like Elastic and Starburst to enable federated analytics across hybrid estates and cyber resilience tools to ensure trust and compliance.
Data management is provided through a composable control plane that unifies governance, pipelines and metadata for adaptable, agent-ready environments. The integration of hybrid control and cross-cloud interoperability allows Dell customers to run workloads where the data lives, and it positions the AI Data Platform as a foundation for hybrid AI.
The current focus on building a data foundation that can deliver unified access, security, metadata intelligence and hybrid orchestration will be a key enabler for future enterprise success. The rise of AI has changed the fundamental way data is managed today and made infrastructure a centerpiece of business strategy.
“AI adoption is accelerating due to the need for real-time insights and automated decision-making, making robust data infrastructure a critical enabler,” Nashawaty said. “The rise of AI-driven applications has placed a greater emphasis on data architecture as a foundational element for AI success.”
