HBM Shortage Drives Memory Chip Prices Up 60-80% in AI Boom

Memory makers capture outsized AI profits

The artificial intelligence infrastructure boom is delivering windfall profits to a small group of memory chip manufacturers as constrained supply of high-bandwidth memory (HBM) drives dramatic price increases across the semiconductor sector.

Micron Technology reported DRAM memory chip prices climbed more than 60% in its quarter ending May 28 compared to the prior quarter, according to the Wall Street Journal. NAND flash memory prices jumped over 80% in the same period. Crucially, shipment volumes increased only modestly, revealing that revenue growth stemmed primarily from price expansion rather than volume gains.

The pricing power reflects a fundamental supply-demand imbalance. HBM chips have become essential components in AI servers and data centers, but manufacturing capacity remains severely limited. Only three companies—Micron, Samsung Electronics, and SK Hynix—dominate global HBM production, and expanding capacity requires multi-billion dollar investments that take several years to complete.

Why it matters

The memory chip shortage represents a critical bottleneck in AI infrastructure scaling. While attention has focused on GPU availability, HBM constraints are now directly impacting product pricing and competitive dynamics across the technology sector. Companies absorbing these costs today may eventually need to pass them to customers, potentially slowing AI adoption or forcing consolidation among smaller players who cannot sustain elevated infrastructure spending.

Costs ripple through the tech sector

Rising memory prices are forcing difficult decisions across the industry. Apple raised prices on multiple MacBook and iPad models this week, explicitly citing increased memory costs as the driver. AI developers face mounting infrastructure expenses that compress margins and complicate unit economics.

Many AI companies have chosen to absorb higher costs rather than raise prices, prioritizing user growth and market share over near-term profitability. This strategy works while venture capital remains available, but creates vulnerability if funding conditions tighten.

Cloud providers including Microsoft, Amazon, Alphabet, and Meta Platforms continue heavy AI infrastructure investments despite elevated component costs, betting that scale advantages will eventually offset higher input prices.

Investor attention shifts to semiconductor suppliers

Equity markets have recognized the profit shift. Shares of Micron, SK Hynix, and Samsung have outperformed many larger technology companies this year as investors price in sustained pricing power for memory manufacturers.

The semiconductor industry is experiencing one of its strongest pricing environments in years, with memory makers positioned as key beneficiaries of AI infrastructure spending rather than just component suppliers.

Additional manufacturing capacity is expected online over the next several years, though current supply constraints will likely persist until new facilities reach production. The timeline for relief remains uncertain, leaving memory manufacturers in a favorable negotiating position for the foreseeable future.

These details were first reported by the Wall Street Journal.