One of the world’s largest memory chipmakers is signaling that the global shortage of RAM and storage is no longer a cyclical blip but a prolonged structural challenge, as the explosive growth of artificial intelligence continues to outstrip supply across the semiconductor industry.
Micron Technology CEO Sanjay Mehrotra delivered the warning during the company’s latest earnings call, telling investors that tight conditions across memory markets are likely to persist for years rather than quarters. The imbalance, he said, reflects a sharp acceleration in AI-related demand colliding with the hard physical limits of semiconductor manufacturing.
“Sustained and strong industry demand, along with supply constraints, are contributing to tight market conditions and we expect these conditions to persist beyond calendar 2026,” Mehrotra said, tempering expectations that new capacity will quickly ease shortages.
At the heart of the supply crunch is the scale of AI data center expansion. Mehrotra said customers have significantly upgraded their buildout plans in recent months, forcing Micron and its peers to revise demand forecasts sharply upward.
“Over the last few months, our customers’ AI data center buildout plans have driven a sharp increase in demand forecast for memory and storage,” he said. “We believe that the aggregate industry supply will remain substantially short of the demand for the foreseeable future.”
AI workloads consume vastly more memory than traditional computing. Training and running large language models requires high-bandwidth memory, advanced DRAM, and large volumes of NAND storage, often deployed in dense, power-hungry server configurations. As AI adoption spreads from hyperscalers to enterprises and governments, memory intensity per server continues to rise, compounding pressure on supply.
Mehrotra said these dynamics are affecting both major memory segments. “Together, these demand and supply factors are driving tight industry conditions across DRAM and NAND and we expect tightness to persist through and beyond calendar 2026,” he noted.
Supply growth constrained by long lead times
While demand has surged, the industry’s ability to respond remains limited. Building new semiconductor fabs or expanding existing ones is a multiyear process that requires massive capital investment, access to advanced lithography tools, and a highly specialized workforce.
Even with aggressive spending, new facilities can take several years before reaching meaningful output. Advanced memory products such as high-bandwidth memory add another layer of complexity, with lower yields and more intricate manufacturing steps than conventional DRAM.
Mehrotra acknowledged that Micron is already operating under strain.
“Despite significant efforts, we are disappointed to be unable to meet demand from our customers across all market segments,” he said, underscoring that shortages are not confined to AI alone but are spilling into other end markets.
Industry discipline reinforces tightness
Micron’s outlook aligns with recent signals from its main competitors, Samsung Electronics and SK Hynix, both of which have cautioned that memory constraints are likely to linger. Rather than rushing to add capacity, leading manufacturers are exercising restraint, shaped by hard lessons from past boom-and-bust cycles.
Samsung has previously said it prefers maintaining long-term profitability and balance sheet strength over aggressive expansion that could trigger another oversupply crash. SK Hynix has similarly highlighted the risks of overbuilding, particularly given the capital intensity and technical difficulty of scaling advanced memory for AI accelerators.
This discipline means that even as prices rise, supply growth will remain measured, reinforcing Mehrotra’s view that shortages are structural rather than temporary.
Broader implications for tech and pricing
For customers, prolonged tightness in memory markets is likely to keep prices elevated, raising the cost of AI infrastructure and pressuring margins for cloud providers, hardware makers, and enterprise IT budgets. Consumer electronics manufacturers could also face higher component costs, especially as devices increasingly rely on larger memory configurations to support AI features.
For memory producers, however, the environment supports stronger pricing power and improved profitability, provided they can execute on advanced products and manage capacity carefully.
Mehrotra’s comments point to a broader shift underway in the semiconductor industry. The AI boom has fundamentally altered demand patterns for memory, and the supply side is structurally constrained from responding at the same pace. Rather than a short-lived shortage, Micron is warning of a new normal in which tight memory markets extend deep into the second half of the decade.
