SK Hynix Delivers Record Quarter as AI Memory Demand Outpaces Supply
Table of Contents
You might want to know
- What drove SK Hynix’s record third-quarter revenue and operating profit?
- Why does the company expect memory supply to remain tight through 2030?
Record results and the AI memory squeeze
SK Hynix reported record third-quarter results on Wednesday, with revenue of 104.6 trillion won (almost $78 billion) and operating profit of 83.3 trillion won ($65.5 billion), according to its earnings release. It was the first time the South Korean memory-chip maker’s quarterly sales exceeded 100 trillion won ($74.5 billion). The figures reflect intense demand for memory used in artificial intelligence infrastructure, particularly as major technology companies and other organizations expand their computing capacity.
Revenue rose 328% from a year earlier, while operating profit increased 632%. The resulting operating margin was 80%: in practical terms, the company earned about 80 won in operating profit for every 100 won of sales. Analysts surveyed by Bloomberg had expected operating profit of 77.5 trillion won, meaning the reported figure was higher than that estimate. The combination of rising sales and a particularly strong operating margin illustrates how favorable supply-and-demand conditions for advanced memory have contributed to the quarter’s performance.
Net income was 39.8 trillion won, up 216% from a year ago but down 58% from the second quarter. The difference between operating profit and net income was affected by costs and losses outside the company’s core chip business. Chief financial officer Kim Woo-hyun attributed the gap to net non-operating expenses of 32.6 trillion won. These included an 11.6 trillion won currency loss and a 22.9 trillion won valuation loss on investments. Those figures help distinguish the strength of the company’s operating performance from the impact of financial items recorded beyond chip sales.
SK Hynix told investors that memory demand is likely to exceed supply through 2030. The company also said its DRAM and high-bandwidth memory (HBM) volumes through 2029 are effectively sold out. DRAM is the standard memory used in servers, phones and PCs. HBM is a stacked type of memory positioned alongside AI processors, where it supplies data for processing. Demand for HBM has grown as businesses build and deploy systems that require large volumes of data to move quickly between memory and processors.
Kim said the industry’s ability to add supply is constrained by the complexity of advanced manufacturing and the time needed to build new facilities. “Given the complexities involved in implementing leading-edge processes and the lead times required to construct new production facilities,” he said, “a situation where supply cannot keep up with demand is expected to persist for some time, a constraint projected to continue through 2030.” The company’s outlook therefore reflects not just current demand, but also the limits on how quickly production capacity can be expanded.
SK Hynix also described signs that customers are adapting their product plans to the shortage. Some customers have recently adjusted their HBM specifications, which the company interprets as evidence that demand is overwhelming available supply. When customers cannot obtain all the chips they want, they may change the specifications of the systems they are building so they can make more AI accelerators from a limited quantity of memory. On the company’s call, SK Hynix said, “Demand remains so strong that customers’ volume requirement is more than double that of what we can supply today.”
To secure future demand and provide greater visibility into orders, SK Hynix has signed long-term agreements with 14 key customers, including leading global cloud providers. These contracts reserve chip volumes years in advance and cover more than 50% of the company’s 2027 and 2028 production capacity. They typically run five years, allow prices to move within an agreed range and include safeguards such as deposits. Such arrangements give customers a way to secure supply over time while giving SK Hynix greater certainty about planned production and commercial commitments.
The company expects investment to rise significantly in 2027, when the first phase of its Yongin chip complex enters full-scale operation. The timing is relevant to its longer-term supply outlook: new manufacturing capacity takes time to establish, and the company has warned that the demand-supply imbalance may remain in place for years. The planned investment is part of the effort to support production, although the company’s comments indicate that capacity additions will not immediately remove the constraints it expects through 2030.
The earnings call also addressed whether continuing AI investment is sustainable when some investors question how quickly spending will generate returns. Citigroup analyst Peter Lee opened the question-and-answer session by asking what supports confidence in AI investment when monetization appears slow. The concern had also been raised eight days earlier by Citadel founder Ken Griffin, who said investors are “almost certainly” in an AI bubble. SK Hynix CEO Kwak Noh-Jung responded by pointing to growth and demand indicators from major cloud customers rather than relying solely on predictions about future AI revenues.
Kwak said major cloud customers reported year-over-year cloud revenue growth of 30% to 80%, while their cloud order backlogs were up two-fold to five-fold. He also said AI spending is broadening beyond a small group of large technology companies. Demand, in the company’s view, is expanding to non-tech firms, sovereign AI projects—government-backed national computing—and on-premises setups, where companies run AI on their own servers. A broader set of buyers could mean that a slowdown by one customer or customer group would have a less significant effect on overall demand.
However, access to financing is another factor shaping the pace of AI infrastructure development. On Oct. 25, Elon Musk wrote that “debt markets have almost entirely shut off access to money” for firms without large non-AI revenue. This comment highlights a potential difference between demand for computing resources and the ability of every infrastructure company to fund new projects. SK Hynix’s outlook is based on demand signals from customers, while financing conditions could still affect how quickly some buyers build capacity.
The outlook for traditional electronics markets is less positive. SK Hynix expects weaker demand from PC and smartphone makers because memory prices are higher and supply is tight. The company nevertheless expects demand from AI infrastructure to offset that decline. The balance between these markets will matter: PC and smartphone demand may be sensitive to rising component costs, while AI data centers and other infrastructure projects are currently driving substantial interest in advanced memory.
Overall, the quarter presents two connected stories. First, SK Hynix achieved exceptionally strong financial results, with sales and operating profit reaching record levels. Second, the company expects the supply constraints behind the strength to remain significant, as production expansion takes time and customers continue to seek more memory than the company can currently provide. The long-term agreements, capacity plans and broadening customer base offer signs of how the company is responding, while questions about financing and the eventual returns on AI investment remain part of the wider market discussion.
Key Insights Table
| Aspect | Description |
|---|---|
| Third-quarter revenue | 104.6 trillion won (almost $78 billion), up 328% from a year earlier. |
| Operating profit and margin | 83.3 trillion won ($65.5 billion), up 632%, with an 80% operating margin; analysts surveyed by Bloomberg had expected 77.5 trillion won. |
| Net income | 39.8 trillion won, up 216% year over year but down 58% from the second quarter. |
| Non-operating expenses | 32.6 trillion won, including an 11.6 trillion won currency loss and a 22.9 trillion won valuation loss on investments. |
| Memory supply outlook | The company expects demand to exceed supply through 2030; DRAM and HBM volumes through 2029 are effectively sold out. |
| Customer agreements | Agreements with 14 key customers cover more than 50% of 2027 and 2028 production capacity, typically run five years and include safeguards such as deposits. |
| AI customer indicators | Major cloud customers reported year-over-year cloud revenue growth of 30% to 80% and backlogs up two-fold to five-fold. |
| Investment plans | SK Hynix expects investment to rise significantly in 2027 as the first phase of its Yongin chip complex enters full-scale operation. |
Afterwards...
SK Hynix’s record quarter underscores how central memory has become to AI infrastructure, while its supply outlook suggests that the current imbalance may outlast the immediate investment cycle. The company’s long-term customer agreements and planned investment provide a framework for managing demand and expanding capacity, but the pace of new supply will depend on complex manufacturing projects and long construction lead times. In parallel, broader adoption of AI could diversify demand, while tighter financing and weaker PC and smartphone demand remain potential counterweights. Investors and customers will be watching whether AI demand continues to grow, how quickly additional capacity becomes available, and whether the economics of AI infrastructure support the investment now being made.
Last edited at：2026/10/28
