SK Hynix to Launch Indiana AI Memory Hub by 2029 Amid Projected Shortage
The semiconductor giant warns that high-bandwidth memory constraints will persist through 2030 despite U.S. expansion.
SK Hynix is establishing a new AI chip production hub in Indiana to meet the escalating global demand for artificial intelligence infrastructure. The company expects to begin mass production of next-generation high-bandwidth memory (HBM) at the facility in the second half of 2029.
According to reports from Reuters, the move is a strategic effort to expand the company's manufacturing footprint within the United States. This expansion aims to place production closer to key U.S.-based customers and mitigate geopolitical risks associated with concentrated supply chains. However, the timeline for the new plant arrives alongside a sobering outlook for the industry: SK Hynix Chairman Chey Tae-won has forecasted that the AI-driven memory shortage will persist through 2030.
The HBM Bottleneck
The current surge in generative AI has created an unprecedented reliance on High Bandwidth Memory. HBM is a critical component for AI accelerators, such as those produced by NVIDIA, as it allows the massive amounts of data required by large language models to move quickly between memory and the processor. Because HBM is significantly more complex to manufacture than standard DRAM, supply has struggled to keep pace with the rapid deployment of AI data centers.
Industry Implications
The projection of a shortage lasting until 2030 suggests that the global semiconductor supply chain remains under extreme pressure. For the broader tech industry, this indicates that the scaling of AI infrastructure may be constrained by memory availability for several years. If HBM remains a bottleneck, the pace at which companies can train and deploy larger, more capable AI models could be slowed, potentially limiting the speed of AI innovation across the market.
Future Outlook
As SK Hynix diversifies its production sites, the industry will be watching whether other memory providers can accelerate their own capacity expansions to offset the projected deficit. While the Indiana facility represents a significant long-term investment in U.S. soil, the gap between now and 2029 remains a critical window of vulnerability for AI hardware availability. It remains to be seen if alternative memory architectures or breakthroughs in manufacturing efficiency can shorten the projected shortage timeline.