Enterprise AI Requires Unified Data Architecture Across Edge and Cloud, Says Pure Storage VP

For years, cloud migration was the default playbook for scaling enterprise infrastructure. But in the age of AI and hybrid work, that strategy is evolving. Amid the market shifts triggered by the Broadcom-VMware acquisition and the changing economics of public cloud at scale, organizations are reevaluating the "move everything" mandate in favor of a more flexible approach.

According to Nirav Sheth, VP and global leader of Sales and Customer Success Engineering at Pure Storage, the solution is a unified data architecture. With experience managing multi-billion-dollar revenue streams and global teams at tech giants like Google Cloud, Okta, and Cisco, Sheth has seen the market from nearly every angle. From his perspective, the future belongs to leaders who can execute a "cloud-smart" strategy.

“Ultimately, any decision, whether a business decision or one feeding an AI investment, starts and ends with data," Sheth says. He cites a recent experience navigating tariff uncertainty as a prime example. To protect customers from financial impacts, Pure Storage needed to analyze data across demand forecasting, supply chain, and manufacturing models. Because the company had already integrated those data sources, it could model scenarios instantly.

Achieving this level of visibility is the key step toward unlocking the full value of AI, Sheth explains. “The right answer is an ‘and,’ not an ‘or.’ Some workloads make sense in the cloud, and some workloads make sense on-prem. Our goal is to deliver value no matter where the data resides.” From his perspective, two significant external forces are driving this shift toward hybrid architectures:

• Geography of risk: “We're seeing specific industries and organizations getting targeted because of the global geopolitical climate,” Sheth notes. He points to a study commissioned by Pure Storage in Australia that highlights a critical trend: organizations that once had a "flexible mindset" toward data sovereignty now prioritize keeping data "in-market" and "in-region" to mitigate risk.
• Power struggle: Citing a recent Federal Reserve study, Sheth explains that kilowatt-per-hour rates in the U.S. have risen 30-40% over the last decade. “We have cities across the US that are literally declaring they can no longer support new data centers because they don't have the power or anything else to sustain them."

The situation is significant enough that hyperscalers are signing custom contracts with nuclear power facilities—"investing to fire up Three Mile Island"—just to secure energy. In this environment, infrastructure efficiency is becoming a core operational metric.

• Bridge the divide: To solve the "And, not Or" equation, Sheth advocates for an Enterprise Data Cloud that functions consistently across the edge, the data center, and the public cloud. But the key enabler is an "intelligent control plane," which allows organizations to decouple governance policies from the underlying infrastructure. "With the old model, every time a new database volume was created, you’d have to reapply policy manually."
• Set it and forget it: "The new model, with our intelligent control plane, is automatic. There’s no risk of human error where a database is suddenly exposed." By automating these policies, small IT teams can manage exabytes of data across a sprawling hybrid estate, Sheth explains. This provides the flexibility to move workloads based on business logic—refactoring apps for the cloud or bringing heavy workloads back on-prem—without sacrificing security or governance.

Navigating this architectural journey requires a shift in the vendor-client relationship, Sheth says. Here, he's candid about Pure Storage’s evolution from a sales-driven culture to a partnership model. "If I think back to ten years ago, our attitude was just to sell whatever we can. Today, that includes being willing to say that a particular solution isn't the right fit and that they would be better off working with someone else," he says. "That’s how you build trust."

For executives facing the complexity of hybrid cloud and the opportunities of AI, Sheth offers a final piece of advice: focus on the blueprint first. “Start with your strategy, map your current state versus where you want to go, and then apply the right technology,” he concludes. “That’s how you build a governance framework that works in the AI era.”