The global RAM shortage: what technology leaders need to know

The world’s appetite for artificial intelligence has triggered a supply chain crisis unlike anything we’ve seen since the pandemic. Memory chips, those unassuming components that quietly power everything from smartphones to hyperscale datacentres, are now at the centre of a global scramble. Prices are soaring, inventories are collapsing, and procurement teams are feeling stuck.

What’s driving the shortage?

The surge in AI adoption is the primary culprit. Training large language models and running inference at scale requires enormous volumes of high-bandwidth memory (HBM). These chips feed graphics processing units (GPUs) in datacentres, enabling the performance leaps that underpin today’s AI boom.

To meet this demand, manufacturers like Samsung, SK Hynix, and Micron have diverted production capacity away from conventional DRAM and NAND flash, the memory that powers laptops, smartphones, and enterprise servers. The result? A perfect storm:

• Inventory collapse: DRAM stock levels fell from 13–17 weeks in late 2024 to just 2–4 weeks by October 2025.
• Price shock: Some memory segments have more than doubled in cost since February.
• Long-term squeeze: Analysts warn shortages could persist until 2027, with 2026 production already sold out.

This isn’t just a semiconductor story. It’s a macroeconomic risk. Rising component costs are pushing up device prices, slowing AI-driven productivity gains, and adding inflationary pressure at a time when global economies are already fragile.

Who’s feeling the pain?

• Tech giants: Hyperscalers are negotiating open-ended supply agreements, effectively saying, “We’ll take whatever you can make.”
• Consumer electronics: Smartphone makers are warning of price hikes of up to 30%.
• Enterprise IT: Servers, storage arrays, and networking gear are all exposed. Procurement teams face unpredictable lead times and escalating costs.

Why it matters for Technology Leaders

For CIOs, CTOs, and Heads of Procurement, this shortage isn’t a passing inconvenience, it’s a strategic inflection point. Memory is foundational to digital infrastructure. Without it, AI projects stall, cloud migrations slow, and device refresh cycles become prohibitively expensive.

What can you do?

Here are five practical steps to navigate the turbulence:

• Secure Long-Term Agreements
  Negotiate extended contracts with trusted suppliers. Many organisations are moving to multi-year commitments to guarantee allocation.

• Diversify Your Supply Chain
  Avoid single-region dependency. Explore secondary suppliers and consider nearshoring strategies to reduce geopolitical risk.
• Optimise Existing Assets
  Extend lifecycle management for servers and endpoints. Invest in memory optimisation tools to squeeze more performance from current hardware.
• Forecast Aggressively
  Use AI-driven demand planning to model scenarios and anticipate allocation challenges. Visibility is now a competitive advantage.
• Collaborate with Partners
  Engage with technology partners who can provide market intelligence, sourcing options, and flexible financing models to mitigate cost volatility.

The bottom line

The AI revolution is rewriting the rules of the global supply chain. Memory chips, once a commodity, are now a strategic asset. Organisations that act decisively will not only weather the storm but position themselves for growth in an AI-driven economy.

At Softcat, we’re helping customers rethink procurement strategies, optimise infrastructure, and build resilience into their supply chains. If you’d like to explore how we can support your organisation, get in touch directly with your Account Manager, or Sales@softcat.com.