Franklin AI Explainer

SK hynix reports record Q3 results as AI memory demand grows

Key Takeaways

  • SK hynix’s official Q3 release reports 632% annual operating-profit growth, alongside stronger server-memory and HBM4 sales.
  • SK hynix reported record third-quarter results in its October 28 earnings release.
  • Revenue reached 104.6155 trillion won and operating profit reached 83.3338 trillion won.
  • The company linked the results to demand for AI memory and its higher-value server products.
  • What the official numbers show The earnings table reports operating-profit growth of 632% compared with the third quarter of 2025.

SK hynix reported record third-quarter results in its October 28 earnings release. Revenue reached 104.6155 trillion won and operating profit reached 83.3338 trillion won. The company linked the results to demand for AI memory and its higher-value server products.

What the official numbers show

The earnings table reports operating-profit growth of 632% compared with the third quarter of 2025. That is the figure in the company's release, rather than the roughly 700% increase repeated in some secondary coverage.

The distinction is worth preserving because a large increase can be distorted by rounding or by mixing different reporting periods. The same release reports quarter-on-quarter and year-on-year changes separately. Those comparisons answer different questions about the business.

What SK hynix says is driving demand

The company describes expansion in server DDR5 sales, HBM4 shipments and high-value-added products for data centres. It also discusses production plans and long-term customer agreements as part of its outlook.

Those forward-looking statements should remain separate from the results already reported. Plans to expand capacity, expectations about future product mix and forecasts for market supply are management assessments, rather than completed outcomes established by the quarter's accounts.

How to read the AI infrastructure signal

For readers following AI infrastructure, the release gives a supplier-level view of how demand is affecting memory revenue and production priorities. It does not measure the usefulness or profitability of every AI application buying that infrastructure.

Keep the currency, reporting period and measure attached to each number. Revenue, operating profit and net income are different items, and a percentage increase does not substitute for the underlying amounts. The release also describes the financial results as preliminary and subject to the review process. Use that scope when comparing this report with subsequent company disclosures.

Our read

Franklin AI Take

The official earnings table gives 632% year-on-year operating-profit growth. Using that exact figure and reporting period is more useful than amplifying a rounded headline. For AI infrastructure analysis, the next question is how the company’s stated memory production priorities develop, while keeping future plans distinct from revenue and profit already reported.