Samsung Electronics is taking a massive leap across the semiconductor industry. In Semiconductor, they are especially focused in the area of Dynamic Random Access Memory (DRAM) production. Through the expansion, the company aims to turn on commercial production at a new manufacturing facility in Pyeongtaek, South Korea, by 2028. This expansion announcement coincides with a massive increase in demand for DRAM, especially for use in AI-powered data centers. This enormous demand has further exacerbated existing supply issues and sent prices skyrocketing, making it difficult for businesses who count on these essential parts to stay afloat.
At the Samsung Tech Day in 2024, Samsung demonstrated its ability to manufacture a 16-high stack of DRAM dies using hybrid bonding technology. This milestone creates new avenues for design and manufacturing of memory chips. Samsung has signaled that it’s technically possible for it to manufacture a 20-die stack. This development is enabling new breakthroughs once thought impossible in DRAM technology.
The demand for such advanced memory solutions is at an all-time high. The profitability of DRAM types used to power Graphics Processing Units (GPUs) and other accelerators in AI data centers has reached unprecedented levels. Demand has increased dramatically, but supply-side issues are driving prices up through the roof. That reality is compounded as memory supplies are redirected for other purposes.
The Evolving Landscape of DRAM Technology
The current specification for High Bandwidth Memory (HBM4, now known as HBM) supports a maximum of 16 stacked DRAM dies. Today’s chips use 12 dies at most. As an example, the recently introduced B300 model uses eight of these HBM chips, with each chip using stacks of 12 DRAM dies. All in all, this innovation is an assertion of Samsung’s dedication to increasing performance without compromising on the ever-increasing requirements of AI-powered apps.
The DRAM industry amidst a once-in-a-generation boom-bust cycle. At the same time, it has a rare opportunity to act on that recognition, made possible by an unprecedented build out of AI hardware infrastructure. This convergence of these opportunities and challenges has created a unique opportunity for semiconductor manufacturers.
Economists note that there are two primary strategies to tackle supply issues within the DRAM sector: innovation and the construction of new fabrication facilities, commonly known as fabs. Each alternative poses different degrees of challenges and timelines.
“There are two ways to address supply issues with DRAM: with innovation or with building more fabs,” – Mina Kim, an economist with Mkecon Insights.
Constructing a new fab requires incredible time and resources. Thus, it sometimes can take 18 months or longer to be fully operational. Combined with the growing urgency to keep up with the demands of emerging AI applications, manufacturers are under heavy pressure to ramp up their production capacities.
Market Dynamics and Economic Implications
The economic environment around DRAM is changing drastically and quickly. Micron Technology, one of the three companies that dominate industry production, had previously announced ambitious production goals. Now, instead of 2030, they plan to reach those targets by 2028! This turbocharged expansion is indicative of the increased demand caused by the rapid pace of AI development and the resultant need for AI-specific hardware.
NVIDIA’s data center business exemplifies this trend. Revenue surged up from just under $1 billion in the last quarter of 2019. By the quarter ending October 2025, it had ballooned to a breathtaking $51 billion. Such bombastic growth doesn’t just highlight the importance that DRAM has on supporting AI workloads and data processing capabilities overall.
Meanwhile, companies are investing billions into servers, data storage and network equipment to deliver their applications. With that, the financial stakes continue to increase. Together with a recent allocation of $3.3 billion toward these sectors, the commitment demonstrates the industry’s intentions to continue scaling up its infrastructure.
“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 last statement especially sums up the conundrum the market faces as booming demand keeps pressure on supply chains and causing prices to soar.
Future Outlook for DRAM Production
Even with the hill to climb, there is hope among industry leaders that a homegrown solution is possible. Shawn DuBravac, the chief economist for the Global Electronics Association, featured a number of bright spots’ solutions to today’s supply crisis. Rapid, incremental capacity expansions by existing DRAM leaders, improved yields in advanced packaging technologies, and a deeper, broader diversification of supply chains are all essential ingredients to deliver relief.
As firms like Intel have recently cautioned, meaningful relief is still years out.
“There’s no relief until 2028,” – Intel CEO Lip-Bu Tan.
On the horizon are exciting developments in manufacturing technology and production capacity. We may not see easy answers on the horizon just yet.
As Samsung increases the pressure further with new fabs and cutting edge processes, the DRAM market is due for some serious upheaval. AI’s accelerating demands will lead semiconductor production to the next evolution. This intersection will be one of the most powerful forces shaping our industry in the years to come.