Beijing: A worldwide surge in artificial-intelligence development has unleashed a severe shortage of memory chips, triggering what experts describe as a “new global supply-chain crisis.” With demand for advanced AI servers, high-performance devices, and data-centre infrastructure at an all-time high, the sudden squeeze on memory components is now reverberating across international markets.
Industry leaders report that the boom in AI adoption has pushed manufacturers to prioritise high-bandwidth memory (HBM) chips a key component for AI training over traditional DRAM used in everyday electronics such as smartphones, tablets, and PCs. As a result, inventory levels for conventional memory have collapsed to historic lows, slipping from several months of stock in 2024 to just a few weeks by late 2025.
The scramble for supply has led to a dramatic rise in prices. Some memory products have doubled in cost within the year, leaving manufacturers struggling to maintain production schedules. Electronics retailers in several countries have already begun rationing memory modules, while some smartphone and laptop makers warn that consumer device prices may rise sharply in the coming months.
Technology giants such as Microsoft, Google, Amazon, and Asian AI firms are at the forefront of the buying rush, signing long-term contracts to secure priority access to HBM. This aggressive procurement has intensified the global imbalance, leaving smaller manufacturers and device makers with unpredictable delivery timelines and escalating procurement costs.
Analysts caution that the crisis extends far beyond consumer electronics. Data-centre expansion plans, AI-cloud rollouts, and even national digital-infrastructure projects could face delays due to constrained chip availability. The imbalance has also begun to fuel inflationary pressure, as rising hardware costs trickle down to both consumers and businesses.
Chipmakers including Samsung Electronics, SK Hynix, and Micron Technology have announced plans to expand production capacity. However, industry experts note that scaling up fabrication plants can take years, and meaningful relief may not arrive until 2027 or later. Meanwhile, supply chains worldwide are being forced to reconfigure sourcing strategies, stockpile critical components, and renegotiate supplier contracts.
The ongoing crisis underscores a deeper vulnerability in the global technology ecosystem. While AI innovation continues at a rapid pace, manufacturing capacity especially for memory components has failed to keep up. The resulting mismatch between technological ambition and hardware availability has now become a structural challenge for the global economy.
As countries and companies compete for limited supply, the world is witnessing a pivotal moment: a reminder that even the most advanced digital revolutions depend on physical components and that when production falters, innovation everywhere feels the impact.