SK hynix CEO Kwak Noh-Jung, a key figure in the global AI semiconductor industry, is scheduled to meet Microsoft co-founder Bill Gates and CEO Satya Nadella this week. The high-profile discussions will take place at the exclusive Microsoft CEO Summit 2026, a closed-door event held in Redmond, Washington.
Industry sources confirmed Tuesday that Kwak’s participation underscores SK hynix’s pivotal role in Microsoft’s evolving AI infrastructure strategy. This strategic alignment aims to reduce reliance on Nvidia and bolster Microsoft’s custom AI chip development.
The invitation-only summit, running from Tuesday to Thursday at Microsoft’s headquarters, convenes approximately 100 top global executives and policymakers. Critical topics like generative AI, advanced cloud infrastructure, and the future of AI data centers will be at the forefront of private discussions. Kwak, a returning attendee from the 2024 summit, is also expected to attend a formal dinner hosted by Bill Gates at his private residence, an exclusive event recently revived after a period of hiatus.
Alongside Kwak, LG Uplus CEO Hong Beom-sik is reportedly the only other South Korean telecom or ICT executive invited to this prestigious gathering.
The repeated invitation highlights the deep integration of SK hynix within Microsoft’s ambitious custom-silicon strategy. In a landmark announcement this past January, the Korean memory chip giant was named the sole supplier of its cutting-edge fifth-generation High-Bandwidth Memory (HBM3E) for Maia 200. This marks Microsoft’s inaugural in-house AI inference accelerator, designed to power its sophisticated AI models. Each Maia 200 chip incorporates six 12-layer, 36-gigabyte HBM3E stacks, delivering a substantial 216 gigabytes of total capacity with an impressive 7 terabytes per second of bandwidth. This exceptional throughput is essential for ensuring large-scale AI models operate seamlessly without performance bottlenecks.
Microsoft has already initiated the deployment of Maia 200 in its Des Moines, Iowa, data center, with plans for expansion to Phoenix, Arizona. Manufactured using TSMC’s advanced 3-nanometer process, the Maia 200 chip delivers a 30 percent improvement in performance per dollar compared to Microsoft’s previous-generation hardware, a key advantage highlighted during its January unveiling.
Across the industry, major hyperscalers including Google, Amazon Web Services (AWS), and Meta are actively developing their own in-house AI chips. This trend is driven by a collective effort to lessen dependence on Nvidia GPUs, which continue to be characterized by high costs and supply constraints. Consequently, this shift has significantly broadened the customer base for leading Korean memory suppliers, as most custom AI accelerators still fundamentally rely on HBM technology. Notably, SK hynix commanded a dominant 57 percent share of the global HBM market by revenue in the fourth quarter of 2025, according to insights from Counterpoint Research.
Beyond its HBM leadership, SK hynix also serves as a crucial supplier of DRAM and NAND flash memory to Microsoft, provides HBM for Nvidia’s high-performance GPUs, and is actively expanding its strategic partnerships with other tech giants like Google and AWS, solidifying its position at the forefront of the AI semiconductor revolution.
mjh
