SK Hynix announced a $38 billion investment to expand its memory chip manufacturing capacity, a move with direct consequences for the global AI supply chain. The capital will fund new fabrication plants in South Korea dedicated to advanced high-bandwidth memory (HBM), the critical component inside Nvidia's GPUs and every major AI accelerator on the market.
For Nvidia, currently trading at $218.99, the expansion matters immediately. The company has faced HBM supply constraints that cap how many GPUs it can ship per quarter. More supply from SK Hynix means fewer bottlenecks and higher shipment volumes — a direct earnings catalyst. Cloud buyers Microsoft, trading at $499.86, and Amazon, at $272.26, also benefit from more stable component availability as they build out AI infrastructure.
SK Hynix's aggressive push intensifies competition with Samsung for HBM market share. Both companies are racing to lock in long-term supply agreements with chipmakers and hyperscalers, and SK Hynix's scale commitment gives it a credible lead. The investment validates the bull case for AI infrastructure spending through the end of the decade and supports revenue growth projections across the chip design, manufacturing and cloud layers of the AI stack.
SK Hynix expects its first new plant to begin mass production by late 2027, with additional facilities coming online thereafter. The company said it will detail its capital expenditure timeline and production ramp-up during its next earnings call, scheduled for October. Those updates will be the key data points for gauging HBM supply trajectory and pricing power heading into 2028.
