SK Hynix Posts ₩104 Trillion Quarter and Still Needs $30 Billion

SK Hynix just booked ₩104 trillion ($78.9 billion) in quarterly revenue, up 328% year on year, and crossed the ₩100 trillion mark for the first time in its history. Net income climbed 216% to ₩39.8 trillion (about $30 billion). The operating margin hit 80%. And yet the company says it still doesn't have enough cash to build a US fab for its Solidigm subsidiary — it figures it's roughly $30 billion short.

That's the strange arithmetic at the heart of SK Hynix's third 2026 quarter. The memory maker is, by its own accounting, effectively debt-free, sitting on a net cash position of ₩98.7 trillion ($74 billion). It's a spending problem, not an earnings problem.

Why did the Solidigm IPO get shelved?

CEO Kwak Noh-Jung explained the abandoned US listing of Solidigm, the company's NAND subsidiary, in direct terms. "The IPO was intended to raise external capital for the expansion of Solidigm's production facilities, and thereby to lighten our own investment burden and broaden our capacity for shareholder returns," Kwak said. "However, after carefully considering recent financial market conditions and other factors, we concluded that proceeding at this time would not be the best way to achieve the originally anticipated outcome. We have therefore discontinued our review."

In plain language: markets weren't offering the price. So the fab bill lands back on SK Hynix's own balance sheet, and that bill is big enough that even a ₩98.7 trillion cash pile doesn't comfortably cover it.

What happens to all the money instead?

The company laid out a four-part spending plan:

• Return most of its one-time Kioxia share gain to shareholders — ₩100 trillion ($75.4 billion) headed their way this year. Because the gain is one-off, it sweetens investors without capping how much of the surging operating profits can go into capacity.
• Expand DRAM output by raising the portion of 10nm-class 6th generation (1c) production.
• Push the M15X fab to maximum output by end of year, specifically to grow HBM4 volume.
• Lift 321-layer NAND above half of domestic production capacity in the fourth quarter, boosting supply of high-performance, high-capacity eSSDs.

Production gets another step up in 2027, when the Yongin Phase 1 plant enters full-scale operation. With domestic and overseas infrastructure investment gaining momentum, the company expects production investment to rise significantly.

How solid is the demand story?

HBM4 became the mainstream product this quarter and should overtake HBM3E in both volume and revenue next quarter. The company has signed 14 LTAs — long-term buying and supply agreements — covering more than half of its memory capacity for 2027 and 2028. Management sees memory demand exceeding supply all the way through 2030.

The rest of the portfolio moved in the same direction. Server DRAM sales on DDR5 ramped significantly, as did LPDDR5 (SOCAMM) for AI servers. TLC SSD sales grew on the latest 321-layer NAND, and development of high-bandwidth flash, the HBF technology, continues.

On demand durability, management described a virtuous cycle: AI-driven productivity gains and new revenue in software development, customer service and search feed greater AI adoption, which in turn drives infrastructure investment. The global AI data center infrastructure market, the company reckons, will keep growing steadily. Short-term growth comes from enterprise and personal AI; the mid-to-long-term driver is physical AI — autonomous driving and robotics.

What did the CEO promise next?

Kwak framed the quarter as evidence of a structural shift, not just a hot cycle. "The customer trust built through our technological competitiveness is leading to long-term contracts, and our business structure is transforming into one with earnings stability and mid-to-long term demand visibility," he said. "We will carry our HBM competitiveness over to next-generation products, and we will also increase the rewards we share with our shareholders as our business outcome and investment result grow."

The open question is the US fab. Solidigm needs a home for expanded production, the IPO route is closed, and SK Hynix — for all its record cash — still needs to find another $30 billion or so before that plant becomes real.