# SK Hynix plans $28B NASDAQ listing, second-largest share sale ever > Source: > Published: 2026-07-08 17:17:31+00:00 # SK Hynix plans $28B NASDAQ listing, second-largest share sale ever South Korea's memory chip giant is betting big on the US market to fund its AI-driven high-bandwidth memory expansion SK Hynix just filed with the SEC for a Nasdaq listing that would raise roughly $28.07 billion, making it the second-largest share sale in history. The only offering to top it was SpaceX’s $85.7 billion listing in June 2026. The South Korean memory chipmaker plans to issue 177.9 million American Depositary Shares at a ratio of 10 ADSs per common share, with pricing set for July 9 and trading expected to begin around July 10. The offering represents approximately 2.5% of the company’s total shares. ## Where the money goes The proceeds are earmarked for expanding SK Hynix’s high-bandwidth memory production capacity, the specialized chips that power the GPUs driving the current AI infrastructure buildout. The company is specifically targeting equipment acquisitions, including ASML’s extreme ultraviolet lithography tools. Institutional investors have already signaled roughly $7 billion in indications of interest. The fundraising target was revised downward from an earlier range of $29 to $29.4 billion, reflecting what the company described as current market dynamics and investor sentiment. ## Why Nasdaq, why now SK Hynix’s decision to dual-list in the US addresses the “Korea discount,” a phenomenon where South Korean equities have historically traded at a discount to their US-listed peers. By listing on Nasdaq, SK Hynix gains direct access to a broader pool of global institutional capital. SK Hynix is the world’s second-largest memory chipmaker and a dominant player in HBM, which has become the bottleneck component in next-generation AI systems. Every major GPU from Nvidia’s latest generation relies on HBM stacks, and SK Hynix has been the primary supplier. The SEC filing landed on July 6, 2026, referencing the July 3 closing price on the Korea Exchange. ## The crypto and compute angle High-bandwidth memory is the same class of technology that powers not just AI training clusters but also the GPUs used in crypto mining operations and distributed computing networks. Decentralized compute networks like Render and Akash are essentially renting out the same GPU hardware that needs HBM to function. A $28 billion capital raise in the semiconductor sector is a significant data point in the continued buildout of compute infrastructure. With $7 billion in early institutional interest against a $28 billion target, the book still needs to fill substantially. Building out chip fabrication capacity takes years, and the $28 billion raised today funds equipment purchases that won’t produce revenue for 18 to 24 months at minimum. **Disclosure:** This article was edited by Editorial Team. For more information on how we create and review content, see our [Editorial Policy](https://cryptobriefing.com/editorial-policy/).