When most enterprises map out their journey onto the AI bandwagon, they think of CPUs first, then the great GPU gold rush. Storage isn’t at the top of their priority list, the unglamorous plumbing storing all the intelligence. Venkat Sitaram, Senior Director and Country Head of Dell Technologies’ Infrastructure Solutions Group in India, wants to flip that picture on its head.
I sat down with him in New Delhi, where Dell announced the India availability of PowerStore Elite – an all-flash memory storage platform – alongside a broad sweep of AI infrastructure innovations. In our conversation, Sitaram kept pointing to a more foundational premise, of how the entire AI conversation has the wrong starting point.
“Storage overhaul is the foundation, as data has become the foundational layer for AI success. And if that storage infrastructure is intelligent, highly automated and cyber resilient, don’t you think it’s a compelling proposition to every enterprise today?”
I pushed back on behalf of the sceptics. What makes storage so critical that ignoring it would be a mistake? Venkat didn’t hedge even a little bit in his response.
“Not thinking about storage would be the biggest mistake. Whether you are using agentic AI or in the early stages, it doesn’t matter. Wherever there is data creation, you need a place to store that safely, securely, protected all the time. For this you need an intelligent infrastructure where storage becomes an integral part,” emphasised Venkat.
Also read: Dell and NVIDIA combine to power NxtGen’s largest India AI factory
“The heart of any tech infrastructure is storage,” explained Venkat. “That’s where all the core operations reside. If that is not well configured or well architected, your business continuity is impacted.”
This is where I tried to play devil’s advocate. Fundamentally, I said, storage is a dumb thing – it’s a hard drive, and a hard drive doesn’t need to be smart. So how does storage actually become intelligent, rather than just enabling an intelligent orchestration layer above it?
Venkat Sitaram reached for the hard drive itself to make his point, and this is where the commercial logic of PowerStore Elite came into focus.
“Imagine there is a 500TB hard drive, and with the PowerStore Elite, the intelligence and the software built in, it is allowing you to use 3 petabytes, six times,” he explained. The consolidation, he argued, is the whole game. “So instead of buying six, you are buying only one, and still saving on operational cost.” That 6:1 data reduction figure is the one Dell is putting front and centre with this launch — the company is marketing PowerStore Elite with an industry-best 6:1 data reduction guarantee and up to 5.8 petabytes of effective capacity in a single 3U appliance.
One of the threads I’d jotted down during his keynote was the shifting nature of datacentre itself, how OEM solutions (like the ones from Dell and its competitors) have evolved from the pre-AI-boom era. I asked him how that demand is translating, and what a datacentre stack from five years ago looks like now.
“Look, the cost of economics has changed with the advent of all the cloud technologies that came in. Today it is AI, but if you were to go back seven years, it was cloud,” Venkat pointed out. Why did people go from on-prem to cloud? “They found that leasing that with that economical model is cheaper and running it as an operational cost. So the fundamentals of that economics changed.”
When I asked for a concrete sense of how a server stack has physically evolved or shrunk, Venkat mentioned key numbers to show how the needle has moved: “If you were to run let’s say 100 odd servers with a power consumption of 100 kilowatt. Now within the same power envelope, you can run almost 3x capacity. Imagine 300 servers with the same envelope. And density has only gotten more and more efficient. That’s the change that we’re bringing,” he emphasised.
Also read: Zero tax, $200 billion dreams: India wants to power world’s AI infrastructure
I told him “AI factory” sounds like jargon to anyone not already steeped in this world, and asked for the simplest possible definition. Venkat bristled in response, “How can it be a jargon? It’s not jargon, it’s real,” he said without mincing any words.
“It’s a framework where you bring together your use cases, data, ecosystem and AI accelerated infrastructure with data protection and sustainability in place. Something that can help you run your pilot into production quickly within hours and days is an unimaginable disruption. AI factory as a framework is really catching 5,000 plus customers and growing fast,” mentioned Venkat.
OEMs like Dell are clearly trying to consolidate vertically as much as possible, so is heterogeneity dead in the AI world? How do you marry vertical integration with genuine choice? Dell’s Venkat Sitaram was firm that the two aren’t in conflict.
“Heterogeneity isn’t dead, it will continue to be there. And from Dell’s perspective, it’s about enabling that for our customers. Heterogeneous hypervisors on Dell Private Cloud, heterogeneous GPU solutions on Dell AI factory, whether it is NVIDIA or AMD. Those choices exist to help customers leverage the best economics.”
I asked who the largest customers adopting Dell’s AI Factory solutions in the first wave actually are. As expected, Venkat Sitaram rattled off regulated industries – government, large enterprises, banking, financial services, telco. But his real interest was in where AI capabilities spread next in India.
“I think this is going to proliferate into education and healthcare. Because AI skills have become important,” Venkat suggested. He connected the rise in AI skills directly to infrastructure investment.
“You have to build an infrastructure campus ready for AI. Only then can you produce students who are AI capable and knowledgeable and have industry collaborations, so that they can solve real-world problems. And India has that opportunity and that’s where we’re seeing a lot of investments coming in education. So it’s going to be a pervasive technology across segments. SMEs are also catching up fast.”
Also read: India leads world in AI skill penetration and AI nervousness