×
SK Hynix posts record $6.7B profit as AI chip demand soars
Written by
Published on
Join our daily newsletter for breaking news, product launches and deals, research breakdowns, and other industry-leading AI coverage
Join Now

SK Hynix posted record quarterly profits and announced plans to boost spending this year as the Nvidia supplier works to address investor concerns about slowing AI chipset demand. The South Korean memory chipmaker reported a 69% jump in operating profit to 9.2 trillion won ($6.69 billion) for Q2, driven by strong demand for high-bandwidth memory (HBM) chips essential for AI processing.

What you should know: SK Hynix is doubling down on AI chip production despite market uncertainties, positioning itself as the dominant player in advanced memory technology.

  • The company plans to double HBM chip sales for the full year compared to 2024, though it didn’t quantify its new investment spending for 2025.
  • Revenue climbed 35% to 22.2 trillion won during the quarter, beating analyst expectations of 9.0 trillion won in operating profit.
  • Shares jumped more than 3% in early trading and are up 55% year-to-date, significantly outperforming the broader Korean market’s 32.7% rise.

The big picture: SK Hynix has emerged as the world’s top memory chipmaker, overtaking Samsung Electronics in Q1 due to its leadership in HBM technology crucial for AI applications.

  • The company’s quarterly profit is double what Samsung Electronics expects to report, as Samsung faces a projected 56% plunge in operating profit due to weak AI chip sales.
  • “SK Hynix foresees that increasing competition among big tech companies to enhance inference of AI models would lead to higher demand for high-performance and high-capacity memory products,” the company said in a statement.

In plain English: HBM chips are specialized memory components that work like super-fast storage units for AI systems. When companies like OpenAI or Google train their AI models, these chips help process massive amounts of data quickly—think of them as the high-performance engine that keeps AI running smoothly.

Why this matters: The memory chip industry sits at the center of the AI boom, with companies like SK Hynix supplying the specialized components that enable training and running large language models.

  • HBM chips are essential components in AI chipsets designed by companies like Nvidia, a leading AI hardware manufacturer, that process vast amounts of data for AI model training.
  • The company is in talks with a major customer regarding sales next year which are “proceeding as planned,” though it didn’t elaborate on specifics.

Tariff tensions: U.S. trade policy uncertainty is creating both opportunities and risks for the Korean chipmaker.

  • SK Hynix said earnings were helped by customer demand to increase inventory ahead of potential tariffs after President Trump threatened to introduce semiconductor tariffs.
  • A much-anticipated meeting between U.S. and South Korean officials to discuss tariffs was cancelled, despite hopes rising after Japan and the U.S. reached a tariff deal this week.
  • While SK Hynix said in April that its U.S. export proportion wasn’t high, analysts warn the company could face pricing pressure from customers squeezed by tariffs.

Competitive pressures: Despite record results, SK Hynix faces headwinds from both market dynamics and rival competition.

  • Goldman Sachs, a major investment bank, downgraded the stock to “neutral” last week, expecting HBM prices to decline for the first time next year.
  • The company is bracing for rising competition from rivals in supplying advanced chips to Nvidia, according to analysts.
  • Senior analyst Ryu Young-ho from NH Investment & Securities noted that SK Hynix’s proactive investment response reflects confidence as it faces “major competition from Samsung Electronics.”
Nvidia supplier SK Hynix plans to boost spending after record Q2 profit

Recent News

AI models secretly inherit harmful traits through sterile training data

Mathematical puzzles and numerical data carry invisible behavioral patterns that bypass traditional safety filters.