The memory chip industry is poised for dramatic changes in the coming years. Pioneers such as Samsung and Micron are poised to take the forefront of these thrilling advancements. Samsung expects to begin production at its new facility in Pyeongtaek, South Korea, in 2028. This action is especially critical for addressing the ongoing memory chip shortage. This shift is in direct response to the increasing use of DRAM within graphics processing units (GPUs). With the expansion of AI data centers, the demand for high-performance memory is increasing rapidly.
In recent developments, Samsung researchers successfully demonstrated the ability to produce a 16-high stack with hybrid bonding technologies in 2024, suggesting that even higher capacities could be within reach. The firm suggested that it’s just a configuration away from achieving a 20-die setup. This scientific breakthrough has the potential to dramatically increase performance of memory.
Current Market Landscape
There are currently about 9,000 global, colocation data centers. Excitingly, nearly 2,000 new data centers are planned or actively being built! This expansion exposes the dangerous reality of our increasing reliance on data-driven technologies. It also highlights the pressing need for robust memory solutions to power these advanced systems.
This creates heady growth potential, but as a counter example, companies such as Samsung have been hesitant to add production capacity. In 2023, Samsung reduced its output by half to avoid having prices go below the cost of manufacture. This big-picture choice really illustrates the tightrope that automakers need to walk to keep pace with constantly changing market demands.
It costs a small fortune to build new fabrication facilities (fabs), with costs for state-of-the-art fabs soaring past $15 billion or more. Beyond that, it can take 18 months or more to stand up a new fab and get it online. For companies, this means being very cautious about investing a lot in expansion until they’re assured there won’t be any peaks and valleys.
“There are two ways to address supply issues with DRAM: with innovation or with building more fabs,” – Mina Kim, an economist with Mkecon Insights.
Demand for DRAM and Economic Considerations
Meanwhile, the demand for DRAM that supports GPUs and other accelerators in AI data centers has obviously exploded. This surge is at a historic high. Over the past few years, NVIDIA’s data center business has been a juggernaut. It took off from a bit over a billion dollars in quarterly revenue in Q4 of 2019 to a staggering $51 billion at the end of October 2025. This explosive growth shows the importance of memory solutions that can flow with technology directions and defines the urgency for such memory.
As experts have cautioned, companies are still shy about ramping up their manufacturing capacity. As author Thomas Coughlin reminds us, the average American firm has the capacity to do new expansions just during economic booms. They just don’t have the fiscal space in off years.
“In general, economists find that prices come down much more slowly and reluctantly than they go up. DRAM today is unlikely to be an exception to this general observation, especially given the insatiable demand for compute,” – Mina Kim.
This reluctance to expand is an unfortunate reality that further complicates the already problematic environment surrounding memory chip production. Businesses are contending with a perfect storm of supply chain disruptions, labor shortages and inflationary pressures. They are continuing to drive down cost while meeting increasing demands for service.
Future Expectations and Innovations
Looking at future forecasts, relief from the ongoing memory chip shortage could soon be in sight. For now, all industry experts are convinced that current DRAM titans will execute only modest capacity expansions. They further highlight the need for yield advancements in advanced packaging as well as a wider diversification of supply chains.
Shawn DuBravac, chief economist for the Global Electronics Association, underscores this multi-pronged approach as key to relieving supply pressures.
“Relief will come from a combination of incremental capacity expansions by existing DRAM leaders, yield improvements in [advanced packaging], and a broader diversification of supply chains,” – Shawn DuBravac.
For all these rosy forecasts, Intel CEO Lip-Bu Tan warns that real relief won’t be felt until 2028 at the earliest. His announcement highlights the struggle manufacturers are under to increase production while contending with constant demand pressure.
“There’s no relief until 2028,” – Intel CEO Lip-Bu Tan.
Or take companies such as Micron which are already on track to hit production milestones two years ahead of schedule. This significant move underlines the need for foresight and vision to chart a path through the memory chip industry over the next five years.