The artificial intelligence investment boom is beginning to have a direct impact on the price consumers pay for smartphones and other electronics as intense demand from data centres deepens a global shortage of memory chips.
Manufacturers are increasingly prioritising high-bandwidth memory, or HBM, used in artificial intelligence accelerators because the products are more complex and profitable than conventional memory used in phones, laptops and other consumer devices.
The shift is tightening supplies of conventional DRAM and other memory components, forcing electronics manufacturers to absorb higher costs, reduce specifications or pass the increases on to consumers.
AI Changes the Economics of Memory Chips
Memory has always been a critical component of modern electronics, but generative AI has fundamentally changed demand.
Training and running large artificial intelligence models requires processors capable of rapidly moving enormous volumes of data.
High-bandwidth memory is designed specifically for that task and has consequently become one of the most valuable components in AI servers.
Manufacturers including Samsung Electronics, SK Hynix and Micron — which together control the overwhelming majority of the global memory market — have shifted investment and production capacity towards these higher-margin products.
The result is less manufacturing capacity available for some conventional memory chips.
Some memory components have reportedly increased several times in price as supply struggles to keep pace with demand.
Consumers Begin Paying the Price
The shortage is increasingly affecting companies selling consumer electronics.
Smartphone, computer and device manufacturers have several options when component costs rise: increase retail prices, reduce the amount of memory included in products, accept lower profit margins or cut production.
For premium manufacturers, raising prices can be easier than sacrificing margins or specifications.
The pressure comes as smartphones have already become considerably more expensive at the high end of the market, where manufacturers are adding increasingly sophisticated processors, cameras and on-device AI capabilities.
Ironically, AI is therefore increasing memory requirements at both ends of the technology industry.
Data centres require huge quantities of advanced memory to operate AI models, while manufacturers are simultaneously putting more memory into phones and computers so consumers can run AI features locally.
New Factories Will Take Years
Semiconductor manufacturers are investing billions of dollars in additional production capacity, but building advanced chip factories is neither quick nor simple.
New fabrication plants can take several years to construct, equip and bring into high-volume production.
Major capacity expansions currently under development are expected to begin contributing meaningfully between 2027 and 2030.
Until then, supply could remain tight.
Industry expectations suggest the imbalance between supply and demand may persist until at least 2028 if AI infrastructure spending continues at its current pace.
The development demonstrates how the artificial intelligence boom is spreading beyond software companies and data centres.
Consumers who may never directly purchase an AI service could ultimately pay part of the cost through more expensive smartphones, laptops and other electronics.
