South Korean semiconductor manufacturer SK hynix has warned that the global memory chip shortage could persist until 2030 as growing demand for artificial intelligence (AI) infrastructure puts increasing pressure on supplies. The warning highlights the challenges facing the semiconductor industry as technology companies expand their computing capacity to support increasingly powerful AI systems.
The company issued the outlook alongside its reported financial results for the third quarter of 2026, which showed record revenue and operating profit. Strong demand for advanced memory products has helped SK hynix benefit from the rapid expansion of AI data centres and high-performance computing infrastructure.
According to the provided figures, SK hynix generated approximately $78.4 billion in revenue during the quarter, crossing a major quarterly sales milestone. Operating profit reached around $62.4 billion, while net profit stood at approximately $29.8 billion.
These figures represent an operating profit margin of about 80% and a net profit margin of approximately 38%, based on the supplied financial data. The reported performance underscores the importance of advanced memory chips in the expanding AI market.
One of the main drivers of growth is high-bandwidth memory (HBM), a specialised type of memory designed to transfer large amounts of data quickly between memory and processors. HBM is particularly important for AI accelerators that train and run complex machine-learning models.
As companies invest in AI servers and data centres, demand for high-performance memory has increased. These systems require substantial memory bandwidth to process large datasets efficiently, making advanced memory products a critical component of modern AI infrastructure.
SK hynix has emerged as a major supplier of high-bandwidth memory, including newer generations designed for advanced AI processors. The company said demand for its HBM products continued to strengthen, with HBM4 becoming an increasingly important part of its product portfolio.
HBM4 is the next generation of high-bandwidth memory technology, developed to support demanding computing workloads. Its adoption is expected to be influenced by AI processor development, customer demand, manufacturing capacity, and the availability of compatible systems.
The company expects HBM4 to surpass the older HBM3E standard in both sales revenue and shipment volume during the fourth quarter of 2026, according to the supplied report. If achieved, this shift would highlight the growing importance of next-generation memory products in the AI semiconductor market.
Demand is not limited to high-bandwidth memory. SK hynix has also reported increased demand for server DDR5 memory and LPDDR5 SOCAMM memory modules. These products serve different roles in computing systems, including server infrastructure and memory-efficient platforms designed for demanding workloads.
The wider memory market is being shaped by the expansion of AI infrastructure, which requires substantial quantities of advanced chips and supporting components. When demand grows faster than manufacturing capacity, suppliers can face difficulties meeting orders, potentially extending delivery times and putting pressure on prices.
Expanding semiconductor production is a complex process that requires significant investment, specialised equipment, technical expertise, and time. Manufacturers must also consider future demand before committing resources to additional capacity, as the memory market has historically experienced periods of both shortages and oversupply.
SK hynix’s warning that shortages could continue until 2030 reflects concerns about the long-term balance between supply and demand. However, the precise duration and severity of any shortage will depend on production expansion, technological developments, customer purchasing patterns, and broader economic conditions.
Competition among major memory manufacturers is another important factor. According to Counterpoint Research’s September report cited in the supplied material, SK hynix held approximately 25% of the global DRAM market by revenue. Samsung led with 38%, while Micron accounted for around 24%.
These figures illustrate the competitive nature of the memory semiconductor industry, where manufacturers are investing in advanced technologies to meet changing requirements. DRAM is widely used in computers, servers, smartphones, and other electronic devices, making supply conditions relevant to several technology markets.
A prolonged memory shortage could affect more than AI companies. Manufacturers of smartphones, personal computers, servers, and other electronic products may also face higher component costs if supply remains constrained. The eventual impact on retail prices would depend on production costs, supplier agreements, inventory levels, and competition between device manufacturers.
For consumers, the key question is whether the growing demand for AI hardware will put sustained pressure on the availability and cost of memory components used in everyday electronics. While shortages do not automatically lead to higher retail prices, persistent supply constraints can make it more difficult for manufacturers to control costs.
SK hynix’s latest warning highlights how the AI boom is reshaping the semiconductor industry. With demand for advanced memory continuing to rise, manufacturers face the challenge of expanding production while keeping pace with rapidly evolving computing technologies.