Dell Technologies is acknowledging that infrastructure and storage supply still can’t keep up with agentic AI’s insatiable appetite for resources. The company this week reported a “record” AI backlog, with $95 billion in orders waiting to be filled, alongside quarterly earnings reflecting a more than 50% year-over-year increase in AI demand. The numbers paint a clear picture: the infrastructure crunch that has defined the early AI era is far from over, and enterprises are competing fiercely for every available server, storage system, and networking component.
On an earnings call, Dell COO Jeff Clarke said that supply constraints begin at the most fundamental layers of the data center. They start with servers and storage, he explained, but quickly cascade upward until they span the stack to “just about every product going through a leading node.” The comment captures the depths of the current hardware shortage, which has become a defining challenge for AI-driven digital transformation.
“We are doing everything we can to get more supply,” Clarke said. “In today’s environment, that’s a very difficult task.” His remarks suggest that this is not a problem Dell can simply spend its way out of. Rather, the bottlenecks are deeply rooted in the global semiconductor supply chain, from the latest leading-edge fabrication nodes to basic memory controllers and power management chips.
A glimpse of infrastructure demands ahead
Dell reported that, in its financial quarter ending July 31, its revenue was $47 billion, reflecting 58% year-over-year growth. Moreover, revenue in its Dell Infrastructure Solutions Group (ISG) increased 89% to a record $31.8 billion. This massive acceleration was driven primarily by servers configured for AI workloads, including both GPU-accelerated systems and traditional CPU-based servers that are increasingly being tasked with agentic AI inference and other demanding applications.
Much of this growth is in servers, notably traditional central processing unit (CPU)-based servers that are now pulling double duty. As enterprises deploy agentic AI—systems that can plan, reason, and act autonomously—they require substantial CPU compute capacity to manage workflows, coordinate across tools, and run middleware. Demand is “exceptionally strong” in this area, with earnings up 122% year-over-year. That number reflects a wider industry shift: AI is no longer confined to a handful of hyperscale data centers; it is moving into the enterprise mainstream, where existing IT infrastructure must be refreshed and expanded to keep up.
Perhaps most tellingly when it comes to the ongoing demand, the company booked nearly $61 billion in AI server orders in the three months ending July 31. All told, over the last 12 months, it has inked more than $130 billion in AI server orders. That booking velocity suggests that enterprise buyers are not waiting for prices to fall or lead times to shrink. They are placing large orders now, often with multi-quarter delivery timelines, to ensure they have a place in the queue.
Clarke reported that Dell converted $131.7 billion of demand into orders over the last year, and that demand is broadening across enterprise customers, neoclouds, and sovereign cloud providers. Neoclouds—a new breed of cloud providers focused exclusively on GPU and AI infrastructure—are competing aggressively for capacity. Sovereign cloud providers, meanwhile, are building national AI infrastructure to meet data residency and security requirements. To illustrate his point, he noted that the number of customers using Dell AI Factory, the company’s platform built to support AI workflows, has surpassed 6,500. Of those, 3,300 signed on in the last three quarters. Clarke pointed out that, by contrast, it took the company two years to sign on the first 3,200 after debuting Dell AI Factory in May 2024.
“Agentic demand is reshaping the data center,” Clarke said. Inference is “pure demand in our industry.” In fact, Dell anticipates that 3,600 quadrillion tokens will be in use by 2030, representing an 87x increase from today. Tokens are the fundamental units of data processed by large language models, and their explosive growth underscores how AI inference is becoming a mainstream workload. Further, over that same period, training demand is predicted to grow to 850 zettaflops, a 5x jump. While training has dominated the AI conversation since the advent of ChatGPT, inference is now emerging as the larger and more sustained opportunity.
“Enterprise agentic AI is expected to be the single largest workload by 2028,” Clarke said, and by 2030 will account for 75% of all data center demand. That forecast, if accurate, represents a fundamental reordering of corporate IT. Data centers that once ran predictable mixes of enterprise applications, databases, and virtualization are now being asked to host autonomous agents that can perform thousands of tasks simultaneously—all requiring ultra-low latency, high-throughput data movement, and massive amounts of energy.
Enterprises clamor for traditional servers
Dell is seeing a growing trend of customers requiring “meaningful CPU compute capacity” to support AI and agentic workflows. As evidence of this demand, in just its last two financial quarters, it has generated nearly as much revenue from traditional servers and networking as it has in any prior full year in company history. This is a remarkable statistic when considering that Dell has been one of the world’s largest server vendors for over two decades.
Most of this growth comes from existing customers accelerating their investments in traditional IT environments to refresh, modernize, and bolster performance, efficiency, and resiliency. Dell anticipates “significant and durable” refreshes ahead, and heightened security and resiliency requirements are also increasing demand. The shift is not only about artificial intelligence; it is also about the aging installed base of servers that were deployed during the last major refresh cycle. These systems are now reaching end-of-life, driving enterprises to upgrade to platforms with more cores, faster memory, and built-in AI capabilities.
“AI requires modern, disaggregated architectures that keep data accessible and in motion across compute, storage, and networking,” Clarke noted. It is much more than assembling and delivering components; AI deployments require significant engineering, design, and deployment expertise. Some customer engagements, in fact, require upwards of 50 unique designs as enterprises optimize for workload performance, power, cooling and the data center environment, he claimed. Each customer comes with a different mix of GPUs, CPUs, storage tiers, and networking topologies, making standard, one-size-fits-all infrastructure obsolete.
Enterprises want new servers with more cores, more dynamic random-access memory (DRAM), and more storage. However, the constraints remain the same: “DRAM, DRAM, DRAM, followed by NAND, NAND, NAND [flash memory],” Clarke said. There are “spotty” CPU and disk drive shortages, and constraints all the way down the supply chain, from microcontrollers to drives to transistors. This echoes a broader industrial reality: the world is still wrestling with the post-pandemic supply chain shock, but now AI demand has amplified it to unprecedented levels. Memory suppliers, for example, have shifted production capacity toward high-bandwidth memory for GPUs, squeezing availability of conventional DRAM. Similarly, NAND flash is being consumed at record rates by AI training clusters and enterprise storage systems, even as cloud providers and enterprises build out massive data lakes for AI workloads.
Large enterprises and multinational corporations across the globe “would prefer to have products now if we had the supply,” he said. “We are supply constrained in the sense of what we can build in any given quarter.” This is not a problem that Dell can solve with better forecasting or additional sales staff. It is an issue of physical production capacity, and every server vendor in the industry is facing the same wall.
This has led Dell to plan accordingly and optimize configurations with what “bits and bytes” they do have coming in to maximize outputs, with a focus on “getting it out the door,” Clarke said. There are associated lead times that the company is working through, but they’ve been able to “realize greater shipments.” By working with component suppliers to prioritize high-demand systems and streamline manufacturing, Dell has been able to squeeze every possible unit out of its available supply. Yet the fundamental imbalance persists.
“We’ll continue to focus on trying to get more supply, and take the supply we have and optimize the output,” he said.
Reflecting increased need for storage as enterprises prep, manage, and protect huge volumes of data, Dell has also seen strong growth across its PowerFlex, PowerStore, PowerProtect, and PowerVault products. Storage is becoming mission-critical for AI, not just as a place to park training datasets but as the underlying layer that enables data to be processed, indexed, and served to models in real time. As organizations build retrieval-augmented generation (RAG) pipelines and multi-modal AI applications that access enormous data sets, they are discovering that traditional storage architectures cannot keep up.
“Demand remains broad based; enterprises continue to modernize their storage environments as data growth increases the importance of keeping data available and secure,” Clarke said. In the age of AI, data is both the fuel and the product. Any interruption in access could stall operations and erode competitive advantage, driving enterprises to invest heavily in storage resilience.
How customers respond to shortages
Clarke acknowledged that modernization is driving higher core counts, more DRAM, and more storage. Those configurations “cost more than they did last quarter, and the quarter before, and the quarter before.” This persistent upward pricing pressure has created a challenging environment for customers with fixed IT budgets. A server that once cost $10,000 might now cost $14,000, and adding the memory and storage required for AI workloads can push that figure even higher.
Customers are adjusting to these price increases, he noted, deferring purchases because they are unable to sufficiently flex existing budget dollars. In other cases, enterprises are placing orders further in advance to ensure they have access to constrained supplies. This advance ordering has become a standard practice for many CIOs, who are being asked to plan around lead times that stretch multiple quarters. Unlike previous cycles, where servers could be delivered in a few weeks, today’s AI-optimized systems often require months from order to deployment.
“Large, sophisticated customers are acting, first and foremost,” Clarke said. Some are collaboratively planning with Dell to gain a view of their needs further into the future. This deep engagement allows Dell to allocate supply against known demand but also forces customers to make commitments earlier in their technology cycles. “That is a new phenomenon,” he said. “We are working through this demand environment that’s well ahead of supply, helping customers manage.”
As the AI infrastructure crunch continues, enterprises must weigh their need for immediate capacity against soaring costs and uncertain availability. Those that can commit early and design flexibly may be best positioned to navigate the ongoing shortage, while others are left waiting. The $95 billion backlog at Dell is not merely a corporate metric; it is a signal that the race to build scalable AI infrastructure still has a long way to go.
Source:Network World News