SK Hynix posted record third-quarter financial results on Wednesday, announcing revenue of 104.6 trillion won (approximately $78 billion) and an operating profit of 83.3 trillion won ($65.5 billion). This marks the first time the South Korean memory chip manufacturer’s quarterly sales have surpassed the 100 trillion won ($74.5 billion) threshold.
Financial Performance Highlights
Revenue surged by 328% compared to the same period last year, while operating profit saw an impressive increase of 632%. The company achieved an operating margin of 80%, indicating that for every 100 won of sales, SK Hynix retained approximately 80 won in profit from its core operations. Analysts surveyed by Bloomberg had projected an operating profit of 77.5 trillion won.
Net income for the quarter was 39.8 trillion won, a 216% increase year-over-year. However, this figure represented a 58% decrease from the second quarter. Chief Financial Officer Kim Woo-hyun attributed this sequential decline to net non-operating expenses totaling 32.6 trillion won. These charges, stemming from activities outside of chip sales, were primarily driven by an 11.6 trillion won currency loss and a 22.9 trillion won valuation loss on investments.
Outlook for Memory Demand
SK Hynix informed investors that it anticipates demand for memory chips to continue outpacing supply through 2030. The company stated that its volumes for both DRAM and High Bandwidth Memory (HBM) are effectively sold out through 2029. DRAM serves as the standard memory for servers, smartphones, and personal computers, while HBM, a stacked memory solution, is crucial for AI processors, providing them with high-speed data access.
“Given the complexities involved in implementing leading-edge processes and the lead times required to construct new production facilities,” Kim Woo-hyun explained, “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 pointed to recent adjustments in HBM specifications by some customers as evidence of overwhelming demand relative to supply. These adjustments, SK Hynix suggested, necessitate the construction of more AI accelerators due to the limited availability of chips. “Demand remains so strong that customers’ volume requirement is more than double that of what we can supply today,” the company stated during an investor call.
Strategic Agreements and Future Investments
To secure future demand, SK Hynix has entered into long-term agreements with 14 key customers, including major global cloud providers. These contracts reserve chip volumes years in advance and cover over 50% of the company’s production capacity for 2027 and 2028. The agreements typically span five years, allow for price fluctuations within an agreed range, and include safeguards such as deposits.
Looking ahead, SK Hynix anticipates a significant increase in investment in 2027 as the initial phase of its Yongin chip complex is set to commence full-scale operations.
Addressing AI Investment Confidence
During the question-and-answer session, Citigroup analyst Peter Lee inquired about the confidence in AI investment given the perceived slow monetization, a concern echoed by Citadel founder Ken Griffin, who recently suggested investors might be in an AI bubble. CEO Kwak Noh-Jung responded by highlighting that major cloud customers reported year-over-year cloud revenue growth ranging from 30% to 80%, with cloud order backlogs increasing two-fold to five-fold.
Kwak further elaborated that AI spending is expanding beyond a few large technology companies to include non-tech firms, sovereign AI projects, and on-premise deployments. This diversification, he noted, would lessen the impact of any single buyer’s slowdown.
However, financing for AI infrastructure players is becoming more constrained. Elon Musk recently commented on October 25th that “debt markets have almost entirely shut off access to money” for companies lacking substantial non-AI revenue.
SK Hynix anticipates a softening in demand from PC and smartphone manufacturers due to elevated memory prices and tight supply. Nevertheless, the company expects that the robust demand for AI infrastructure will offset this decline.
