The rapid rise of AI workloads is driving a new wave of infrastructure demand, and memory is at the center of it. In this episode of SHI’s Research Breakdown, host Victoria Barber is joined by SHI colleagues Russ Cantwell, Lead Field CTO, and Steve Troxel, Field CTO, to unpack the implications of the emerging memory shortage and the rising cost of data center infrastructure. Their discussion builds on the Gartner research report Rising Costs Ahead: Managing AI-Driven Price Increases in Data Center Infrastructure, which we featured in the February 20th edition of the Research Roundup LinkedIn newsletter.

Rather than treating the memory shortage as a short-term procurement issue, the conversation explores the deeper structural forces at play. Russ explains how growing demand for high bandwidth memory (HBM) used in AI systems can create ripple effects across the broader memory ecosystem, impacting supply and pricing for technologies like DDR5. Steve adds that organizations cannot simply “buy their way out” of the problem. Instead, IT leaders will need a more nuanced strategy that includes optimizing existing workloads, understanding application-level memory usage, and aligning infrastructure investments with evolving business priorities.

The discussion also highlights how traditional supply chain approaches (especially just-in-time infrastructure procurement) are being tested by the AI boom. Russ argues that organizations may need to rethink how they design IT architectures, focusing less on hyper-efficiency and more on maintaining flexibility and control. This includes strategies such as abstracting applications from specific cloud services, diversifying infrastructure options, and designing systems that can adapt when supply constraints or pricing changes emerge.

Ultimately, the current memory shortage may serve as a catalyst for broader change in enterprise IT strategy. Instead of reacting to price increases, CIOs and infrastructure teams have an opportunity to reassess architecture, prioritize resilience, and ensure they retain the ability to pivot as technology markets shift.

Discussed in this episode:

The AI-driven memory shortage and why data center costs are rising
How high bandwidth memory (HBM) demand affects the entire memory market
Why the current memory crisis may be a “supercycle” rather than a typical memory cycle
Strategies for mitigating supply shortages, including right-sizing workloads and extending hardware lifespan
The risks of relying on just-in-time infrastructure procurement
Why control and flexibility are becoming critical design principles in enterprise IT
How CIOs can rethink infrastructure strategy to navigate supply chain volatility