Dell COO: Agentic AI Demand Could Keep Memory and HDD Shortages Going for Over Five Years

Jeff Clarke, chief operating officer of Dell Technologies, said in a September 2026 meeting with Morgan Stanley that agentic AI is fundamentally different from past hardware cycles. If related AI workloads continue to grow, shortages of key components such as memory and HDDs could last more than five years, rather than the usual several quarters.

Clarke argued that agentic AI decouples cognitive output from human headcount, allowing companies to accomplish more without adding staff, potentially driving 10x to 100x productivity gains. This shift is pushing businesses of all sizes to pursue AI transformation and continuously expand the infrastructure total addressable market. Dell estimates that data center compute capacity will add 200 gigawatts by 2030, while inference token generation will grow 87-fold.

Asked whether inference demand would mainly flow to cloud platforms, Clarke said agentic AI workloads will be deployed in a hybrid manner, with enterprises shifting workloads between public cloud and on-premises environments based on security and cost efficiency. The spread of open-weight models is also expected to boost on-premises infrastructure investment.

Regarding the current decline in server shipments, Clarke attributed it to significant increases in server performance density, noting that Dell's 17th and 18th generation servers can replace up to 13 traditional 14th generation servers. As data centers complete their shift to accelerated computing and agentic AI drives CPU server demand, he expects server shipments to resume growth.

On storage, Clarke pointed out that every agentic AI operation creates ongoing data storage needs, and the large-scale application of KV cache will further increase capacity requirements. He listed supply constraints as his biggest concern, saying Dell is operating on the assumption of multi-year component shortages, particularly for memory and HDDs.