Computer memory is expensive because the chips are hard to manufacture, demand keeps rising, and only a handful of companies control the supply

RAM and storage chips require precision manufacturing at a scale that few factories in the world can handle. Each chip contains billions of transistors packed into a space smaller than your fingernail, and a single defect ruins the whole piece. The factories that make these chips—called fabs—cost billions of dollars to build and operate, and they run at full capacity most of the time. When demand spikes or a fab shuts down for maintenance or upgrades, prices jump because there is nowhere else to buy from.

Three companies—Samsung, SK Hynix, and Micron—make most of the world's memory chips. That concentration means they can raise prices when demand is high, and they have little incentive to build extra factories just to lower costs. Geopolitical tensions, trade restrictions, and supply chain disruptions have made this worse in recent years, cutting off some manufacturers from key materials or markets.

Key Takeaways

  • Memory chips are manufactured in only a few factories worldwide, so any disruption to production—maintenance, accidents, or geopolitical conflict—causes when ready price increases.
  • The three largest manufacturers control about 90 percent of the market, which means they can maintain high prices without fear of competition.
  • Building a new memory fab costs $10 billion to $20 billion and takes years, so new factories cannot be built quickly to meet sudden demand.
  • Memory prices fluctuate based on whether the industry is in an oversupply or undersupply cycle, and these cycles can last months or years.
  • Older memory technology (like DDR4 RAM) sometimes costs more than newer technology (like DDR5) because older fabs are less efficient and produce less per dollar spent.

Why there are so few memory manufacturers

Building a memory fab requires informed that took decades to develop. Samsung, SK Hynix, and Micron invested heavily in the 1990s and 2000s, and they have kept their lead by reinvesting profits into research and newer factories. A company starting from scratch today would need to spend $15 billion to $20 billion just to build one fab, then spend years training engineers and working out manufacturing problems before producing a single usable chip.

Intel tried to enter the memory market in the 1980s and 1990s but eventually gave up because the three incumbents had too much of a head start. More recently, Intel announced plans to manufacture memory chips for other companies, but those fabs are still under construction and years away from full production. This means the three-company duopoly is likely to persist for at least another decade.

How manufacturing bottlenecks drive up prices

Memory fabs run 24 hours a day, 7 days a week because the equipment is so expensive that idle time is wasteful. When demand is high—such as during a PC or smartphone boom—every fab runs at maximum capacity and still cannot keep up. Manufacturers cannot straightforward add a second shift or hire more workers; the bottleneck is the physical equipment, not labor.

When a fab needs maintenance or upgrades, it shuts down for weeks or months. During that time, the other fabs have to cover the lost production, but they are already at full capacity. Prices rise because buyers have nowhere else to turn. A fire, accident, or natural disaster at a single fab can ripple across the entire industry for months.

In 2022, for example, a fire at a Samsung fab in South Korea disrupted production for weeks. In 2024, flooding at a Micron fab in Taiwan raised concerns about supply. These events are rare, but when they happen, memory prices spike within days.

The role of supply and demand cycles

The memory industry moves in cycles. When prices are high, manufacturers invest in new equipment and fabs. A few years later, all that new capacity comes online at once, and supply exceeds demand. Prices crash. Manufacturers stop investing, fabs age without upgrades, and supply shrinks. Demand catches up, prices rise again, and the cycle repeats.

These cycles can last three to five years. Right now, the industry is in an undersupply phase following years of strong demand from data centers, AI training, and cloud computing. Prices are elevated, but they are not at historic highs. If you bought memory in 2021 or 2022, you likely paid more than you would today.

Why older memory sometimes costs more than newer memory

DDR4 RAM, which is older technology, sometimes costs more per gigabyte than DDR5, which is newer. This happens because DDR4 is made in older fabs that are less efficient. Those fabs produce fewer chips per dollar of equipment and materials, so the cost per chip is higher. Manufacturers keep running the old fabs because they are paid off and still profitable, but they do not invest in upgrades.

Meanwhile, DDR5 is made in newer fabs with better yields—meaning fewer defects and more usable chips per production run. Over time, as DDR5 fabs mature and older DDR4 fabs are retired, the price gap should close. But during the transition period, older technology can be more expensive.

Geopolitical and trade restrictions

The United States, European Union, and other countries have imposed restrictions on selling advanced memory chips to China and Russia. These restrictions limit the market for manufacturers, which means they have less incentive to build new capacity. They also create uncertainty: a manufacturer might hesitate to invest in a new fab if new trade rules could cut off a major customer.

Taiwan produces a significant share of the world's memory chips, and any disruption to Taiwan—whether political, military, or natural—would be catastrophic for global supply. This geopolitical risk is priced into memory costs; manufacturers charge a premium partly because they know that supply could be disrupted without warning.

What this means for your wallet

Memory prices are unlikely to drop dramatically unless the industry enters an oversupply phase or a new competitor emerges. You can watch prices on sites that track historical trends, and you may find that waiting a few months saves you money if the industry is moving into an oversupply cycle. However, if you need memory now, waiting for a price drop is a gamble.

Buying used or refurbished memory can save 20 to 40 percent, though you lose the warranty. Buying in bulk—if you are building multiple computers—sometimes qualifies you for discounts. And choosing older technology (like DDR4 instead of DDR5) can save money if your use case does not require the newest standard.

Frequently Asked Questions

Will memory prices ever come down?

Yes, but the timing is unpredictable. Prices fall when the industry enters an oversupply phase, which happens when manufacturers have built too much capacity. This cycle repeats every few years, but you cannot predict when it will occur. Watching price tracking websites can help you spot when a drop is beginning.

Why do different brands of RAM cost different amounts if they use the same chips?

Different brands buy chips from the same manufacturers but add their own circuit boards, cooling, and firmware. Some brands test their chips more rigorously or offer longer warranties, which justifies a higher price. Others compete on price and accept lower margins. The underlying chip cost is the same, but the final product price varies based on brand strategy.

Is it cheaper to buy memory in bulk?

Sometimes. Large retailers and system builders negotiate volume discounts with manufacturers. If you are buying 10 or more units, you may may have access to for a discount of 5 to 15 percent. Small purchases at retail prices do not get this benefit, so buying used or waiting for a sale is often more effective for individual buyers.

Could a new company start making memory chips and lower prices?

Theoretically yes, but practically no in the near term. Building a competitive fab costs $15 billion to $20 billion and takes five to seven years. By the time it is operational, market conditions will have changed. Intel is attempting this, but even with decades of chip-making experience and billions in funding, their memory fabs are years away from full production.

Why is memory more expensive in some countries than others?

Import tariffs, shipping costs, local taxes, and currency exchange rates all affect the final price. A country that manufactures memory locally (like South Korea or Taiwan) often has lower prices than a country that imports everything. Trade agreements and restrictions also play a role; countries with trade barriers may pay more.