SK Hynix posted record Q3 revenue of 104.62 trillion won ($75.8 billion USD) and an operating profit of 83.33 trillion won ($60.4 billion USD). The South Korean chipmaker accompanied the earnings spike with a planned 100 trillion won ($72.5 billion USD) shareholder return package for fiscal 2026. However, net profit fell well short of market estimates, landing at 39.53 trillion won ($28.6 billion USD) due to heavy non operating expenses.
Operating profit climbed 7 times compared to the prior year, beating analyst estimates by 9%. Revenue also surpassed Wall Street projections of 97.34 trillion won ($70.5 billion USD). While sales of high bandwidth memory to Nvidia data centers provided strong momentum, surging prices for standard DRAM and NAND chips supplied the main financial lift. DRAM average selling prices rose roughly 20% over the previous quarter, while NAND prices jumped in the mid 20% range. Counterpoint Research research director Hwang MS noted that fast price recoveries across traditional memory lines acted as the primary earnings catalyst.
Net profit missed expectations of 60.73 trillion won ($44 billion USD) by 35%. This divergence between operational earnings and bottom line income came from one time financial deductions. Major factors included foreign exchange losses, tax adjustments tied to the company investment in Japanese flash maker Kioxia, and assorted non operational expenses. Analysts continue reviewing the filings to determine the ongoing impact of these tax liabilities.
SK Hynix allocated 100 trillion won ($72.5 billion USD) for total shareholder returns throughout fiscal 2026. This package directs over 30 trillion won ($21.7 billion USD) toward cash dividends and over 60 trillion won ($43.5 billion USD) to share buybacks and cancellations. The board cleared a Q3 dividend of 20000 won ($14.50 USD) per share, splitting the sum equally between base dividends and special payouts tied to Kioxia returns. Management warned that memory shortages could last until 2030 as production lines shift toward artificial intelligence silicon, which could push up manufacturing costs for consumer smartphones and computers.
