TLDR: SK Hynix has announced a record-breaking third-quarter profit, primarily fueled by the explosive growth in generative AI, which has driven robust sales of high-bandwidth memory (HBM) and increased demand for DRAM chips. The company continues to solidify its leadership in the AI memory market with strategic investments and advancements in next-generation HBM technology.
Seoul, South Korea – September 21, 2025 – SK Hynix, a global leader in semiconductor manufacturing, has reported a record-breaking third-quarter profit, extending its streak of strong financial performance. This unprecedented success is largely attributed to the burgeoning demand for high-bandwidth memory (HBM) and a significant increase in DRAM chip sales, both of which are critical components for the rapidly expanding generative artificial intelligence (AI) sector.
The company’s impressive Q3 results build upon a foundation of robust performance in earlier quarters of 2025. In the first quarter (January-March), SK Hynix posted an operating profit of 7.44 trillion won ($5.4 billion), marking a substantial 157.8% year-on-year increase and representing its second-highest quarterly earnings ever. This momentum continued into the June quarter, where operating income surged by 68%.
Industry analysts point to the “generative AI boom” as the primary catalyst for SK Hynix’s exceptional growth. The increasing adoption of AI models, including more affordable and high-performance options, has spurred aggressive investments from global big tech companies, directly translating into higher demand for advanced memory solutions. SK Hynix is a key supplier to AI giant Nvidia and is widely recognized as a front-runner in the lucrative HBM market.
High-bandwidth memory (HBM) chips are particularly vital for AI accelerators, enabling faster data processing and improved AI service performance. SK Hynix has been at the forefront of HBM innovation, with plans to significantly boost production of its fifth-generation HBM3E chips, aiming for them to constitute over half of its HBM sales in the second quarter. Looking further ahead, the company is on track to complete preparations for mass production of its sixth-generation HBM4 chips by the end of 2025.
In an earnings conference call, SK Hynix CFO Kim Woo-hyun stated, “CAPEX is expected to increase compared to our previous plans, with most of the investment going toward equipment for HBM product”. This strategic focus on HBM is evident in the company’s financial commitments, with R&D spending reaching 3 trillion won in the first half of the year, already exceeding 60% of 2024’s total, and capital expenditure nearly doubling to 11.2 trillion won during the same period.
Justin Kim, president and head of AI Infra at SK Hynix, emphasized the transformative potential of their latest advancements. “HBM4, a symbolic turning point beyond the AI infrastructure limitations, will be a core product for overcoming technological challenges,” he said. Kim added that the company aims to “grow into a full-stack AI memory provider by supplying high-quality, diverse memory products that meet the performance requirements of the AI era in a timely manner”. The company anticipates that HBM4 will improve AI service performance by up to 69%, addressing data bottlenecks and significantly reducing data-center power costs.
Despite global trade uncertainties, including potential U.S. tariffs, SK Hynix has maintained resilient demand for its HBM chips, shielded by long-term supply agreements. Vice President Kim Woo-hyun noted that “Global customers have maintained their planned memory purchases,” downplaying concerns about preemptive stockpiling. The company expects HBM demand to more than double in 2025 compared to the previous year, underscoring the sustained growth in the AI memory market.
Also Read:
- Ibiden Ramps Up IC Substrate Production to Meet Soaring AI Demand
- Arm Foresees Customized Low-Power CPUs as Key to South Korea’s AI Data Center Transformation
SK Hynix’s continued record profits highlight its pivotal role in the AI revolution, driven by its leadership in advanced memory technologies and strategic investments to meet the escalating demands of the global AI industry.
