News · 2026-07-11

SK Hynix's Nasdaq listing raises $26.5 billion, the largest first-time US listing by a foreign company ever

SK Hynix raised $26.5 billion in its Nasdaq debut, the largest first-time US listing by a foreign company in history -- eclipsing the record Alibaba set in 2014. The offering was priced on the back of one thing: SK Hynix makes the high-bandwidth memory that today's AI accelerators cannot function without, and demand for that memory has become one of the tightest bottlenecks in the entire AI buildout.

Key facts

• Amount raised: $26.5 billion, topping Alibaba as the biggest-ever US IPO by a foreign company.
• Pricing: American depositary receipts priced at $149, closed day one up about 13% at $168.
• Demand: roughly seven times oversubscribed; the company sold 177.9 million ADRs (each equal to one-tenth of a share).
• Primary source: financial reporting via Reuters markets coverage.

To see why a memory company just pulled off a record listing, you have to understand what memory does in an AI system. The GPUs everyone talks about -- Nvidia's H100s and B200s -- are the engines, but an engine is useless if you can't feed it fuel fast enough. High-bandwidth memory, or HBM, is that fuel line: stacks of memory chips mounted right next to the processor, connected by a very wide, very fast bus, so the accelerator never sits idle waiting for data. Every large AI training and inference cluster needs enormous amounts of it, and SK Hynix has become the dominant supplier -- the single company most GPU makers depend on for the newest HBM generations.

That dominance is what investors were buying. The offering was about seven times oversubscribed, and the American depositary receipts -- each representing one-tenth of a regular share -- priced at $149 and jumped roughly 13% on the first day of trading to close near $168. Raising $26.5 billion put the listing past Alibaba's decade-old record, a symbolic marker of how thoroughly the AI hardware supply chain has reordered the list of the world's most valuable companies. Where Alibaba's record was about consumer internet scale, SK Hynix's is about being an irreplaceable link in the AI compute chain.

The deeper story is about where the value in AI is accruing. For two years the narrative centered on the labs building models and the GPU makers selling chips. But the buildout runs on a longer supply chain, and the memory makers sitting one layer below the GPU have quietly become just as constrained -- and just as prized. HBM capacity is sold out well into the future, and the companies that make it can name their price. SK Hynix's listing is the clearest financial confirmation yet that the AI infrastructure boom (see the $660 billion buildout) is minting winners well beyond the two or three chip designers that dominate the conversation.

The honest caveat is the one hanging over the whole sector: this valuation assumes the AI capital-spending boom keeps compounding. HBM demand is downstream of GPU demand, which is downstream of the hyperscalers' and neoclouds' willingness to keep buying accelerators at the current pace. If that spending slows -- and there are credible warnings that parts of the buildout are being financed in a circular, self-reinforcing loop -- the memory makers riding the wave would feel it as sharply as anyone. For now, though, the market has spoken with the largest foreign IPO in US history, and it is a bet that the demand for AI memory has a long way to run.

Primary source, verified: read the paper →

Ground Truth. (2026, July 11). SK Hynix's Nasdaq listing raises $26.5 billion, the largest first-time US listing by a foreign company ever. Ground Truth. https://groundtruth.day/news/sk-hynix-largest-foreign-ipo-in-us-history.html

@misc{groundtruth:sk-hynix-largest-foreign-ipo-in-us-history,
  title  = {SK Hynix's Nasdaq listing raises $26.5 billion, the largest first-time US listing by a foreign company ever},
  author = {{Ground Truth}},
  year   = {2026},
  month  = {jul},
  url    = {https://groundtruth.day/news/sk-hynix-largest-foreign-ipo-in-us-history.html}
}

Topics: hardware · memory · ipo · markets · ai-infrastructure