Dell COO warns agent AI could cause memory, HDD shortages lasting over five years

Agentic AI is reshaping the underlying logic of traditional hardware cycles, potentially transforming short-term supply chain volatility into a multi-year structural constraint.

Jeff Clarke, Chief Operating Officer of Dell Technologies (DELL.US), stated in a recent meeting with Morgan Stanley that agentic AI differs significantly from past hardware cycles. If related AI workloads continue to grow, shortages of key components — particularly memory and HDDs, which Clarke cited as his biggest concerns — could persist for more than five years. Unlike past hardware supply crunches that typically lasted several quarters, this wave of demand may create a longer-term supply-demand gap.

Following the meeting, Morgan Stanley reassessed the current AI infrastructure cycle. Analysts noted that the eight hardware boom-and-bust cycles over the past 40 years were largely driven by equipment upgrades and capital expenditure fluctuations. The key variable in this cycle, however, is whether AI workloads can sustain their expansion. In other words, Dell's executive is not describing a typical hardware restocking cycle, but rather a potential long-term infrastructure expansion driven by growing demand for agentic AI.

Agentic AI Reshapes Infrastructure Demand Logic

According to the report, Clarke believes this AI cycle is fundamentally different from the hardware booms and busts of the past several decades. Previous hardware cycles — spanning servers, storage, and PCs — were primarily driven by changes in device penetration rates and upgrade cycles. Demand rose and fell with refresh cycles, but the overall size of the hardware market did not expand substantially.

For example, during the pandemic, PC demand temporarily increased as the human-to-device ratio rose. However, once society reopened and normalized, PC demand receded. The total market size before and after the pandemic remained essentially unchanged.

Agentic AI operates on a different logic. Clarke argues that agentic AI decouples cognitive output from human input. Companies can accomplish more work without adding headcount, resulting in 10x to 100x productivity gains.

In this process, AI is no longer merely a tool assisting humans; it becomes a "productivity engine" that continuously consumes computing resources. This productivity boost will drive more enterprises — regardless of size — to undergo AI transformation, thereby continuously expanding the total addressable market (TAM) for infrastructure.

Dell projects that by 2030:

• Data center computing power will increase by 200 gigawatts.
• ZettaFLOPS computing power will grow 5x to 830.
• Inference token generation will grow exponentially.

Agentic AI requires significantly more tokens per unit of work compared to basic chatbots. Clarke estimates that inference token generation will increase 87x by 2030.

Cloud vs. On-Premises Is Not a Zero-Sum Game

A core question surrounding this AI infrastructure cycle is whether rapidly growing token demand will ultimately flow primarily to cloud platforms rather than on-premises infrastructure.

Clarke's assessment is that agentic AI workloads will adopt a hybrid deployment model. Enterprises will continuously adjust where inference workloads run — between public cloud and on-premises environments — based on security and cost efficiency.

Using Dell as an example, Clarke noted that the company deploys content-related workloads in the public cloud but will not move proprietary source code or telemetry data out of its own facilities. This suggests there is no single deployment path for AI workloads; on-premises infrastructure will continue to absorb a portion of growing computing demand.

Under this framework, exponential growth in token consumption does not necessarily mean cloud and on-premises infrastructure are in a zero-sum relationship. Clarke believes the proliferation of open-weight models will further drive investment in on-premises infrastructure, as enterprises can optimize model output costs in their own environments.

Server Shipments Decline, But Long-Term Demand Logic Remains Intact

Current year-over-year shipments of traditional servers are still declining, which appears to contradict the explosive growth in inference tokens. Clarke explains that significant increases in server performance density are reducing the number of units shipped.

Dell's 17th and 18th generation servers can replace up to 13 14th-generation traditional servers. Therefore, in the short term, even as server shipments decline, the value and capacity per unit continue to rise.

As data centers complete their architectural transformation toward accelerated computing, and as agentic AI adoption further drives demand for CPU servers, server shipments are expected to eventually resume growth. Clarke even suggested that if inference tokens truly achieve 87x growth within five years, traditional server shipments could also experience exponential changes.

Storage will also see sustained demand from agentic AI. Every operation performed by agentic AI — including memory retention and artifact generation — creates ongoing data storage needs. With the large-scale adoption of KV Cache, storage capacity requirements will grow further.

Supply Shortages Could Become a Multi-Year Structural Constraint

Compared to macroeconomic conditions, geopolitical conflicts, data center overbuilding, and power shortages, Clarke listed supply issues as his biggest concern. He described the current supply environment by saying, "We're in neverland."

Morgan Stanley noted that historical commodity supply cycles typically lasted 2 to 4 quarters, driven by boom-and-bust hardware refresh cycles and supply-side decisions. However, Clarke argues that if the growth logic for gigawatts and inference tokens is accepted, key components like memory and HDDs are no longer in a typical short-term cycle.

As a result, Dell is currently operating under the assumption that critical components may face multi-year supply shortages, with particular focus on memory and HDDs. In other words, the core risk of this supply crunch is not short-term insufficiency, but rather that supply may take years to catch up as AI demand continues to expand.

Clarke also stated that Dell's supply management capabilities are superior to its peers, and this advantage is translating into market share gains across multiple segments, including servers, storage, and PCs.

Margin Expansion Reflects Structural Supply-Demand Changes

Persistent tightness in key components is also reshaping the pricing environment in the infrastructure industry.

Morgan Stanley previously attributed part of the gross margin expansion in server and storage businesses to "profit stacking" — where manufacturers add additional profit margins on top of rising component costs. Clarke offered a different explanation.

He acknowledged that margins for similar server and storage products are indeed expanding, but attributed this to two structural factors:

1. When component supply is limited, companies prioritize allocating scarce resources to the highest-margin products.
2. As the infrastructure market continues to expand, the pressure to compete for incremental customers decreases, reducing the need to lower prices to win new business.

Therefore, margin improvement is not entirely opportunistic due to short-term supply-demand mismatches. It also reflects a shift in the industry's pricing benchmark as infrastructure demand expands sustainably.

If the workload growth driven by agentic AI continues to materialize, the core variable of the AI infrastructure cycle will shift from "equipment upgrades" to "new demand." This means the expansion of demand for computing power, servers, and storage could last longer, and supply constraints for key components like memory and HDDs could evolve from short-term cyclical fluctuations into a multi-year structural issue.

This article is republished from Wall Street CN. Edited by Chen Siyu, Zhitong Finance.