Chip Heavyweight Hynix’s Revenue Beats as AI Lifts Memory Demand
Chip Heavyweight Hynix’s Revenue Beats as AI Lifts Memory Demand · Bloomberg

In This Article:

(Bloomberg) -- SK Hynix Inc.’s quarterly sales beat estimates, with the company declaring the beginnings of a recovery in the memory chip market thanks to surging interest in artificial intelligence.

Most Read from Bloomberg

The South Korean supplier to Apple Inc. and Nvidia Corp. now expects sales of its high-end DRAM chips to more than double this year to power storage-hungry AI applications. SK Hynix said it’s gearing up to push out more of those chips, even as the company vowed to continue to slash overall output.

The chipmaker’s sales came to 7.31 trillion won ($5.7 billion) in the June quarter, roughly half its revenue a year ago but soundly beating an average projection for 6.05 trillion won. Its operating loss came to 2.88 trillion won, in line with expectations.

“The worst has passed,” said Baik Gil-hyun, an analyst at Yuanta Securities Korea.

Shares in SK Hynix pared gains and were down around 1.5% Wednesday, amid a broader selloff on the Kospi and with weakness in the overall memory market tempering optimism.

Samsung Electronics Co., SK Hynix and Micron Technology Inc. have struggled for the better part of two years with a post-Covid collapse in demand for the memory chips essential to smartphones, servers and computers. This month, larger rival Samsung recorded its worst decline in quarterly sales in more than a decade, while sector bellwether Taiwan Semiconductor Manufacturing Co. cut its outlook and postponed production at its Arizona project to 2025, underscoring the depth of a global electronics market slump that set in after consumers and corporations trimmed spending to deal with a downturn.

SK Hynix shares recouped some of their losses since 2021 this year, gaining about 51% through Tuesday’s close on bets that output cuts and greater economic certainty might lift industry prices. Analysts including JP Morgan’s JJ Park also cited the rollout of new high-bandwidth memory that works in tandem with AI hardware such as Nvidia graphics accelerators, which promise to speed up data processing for intensive tasks such as training AI.

“We are seeing increasing adoption of AI by major customers, and that makes us believe that a turnaround for memory is very near,” Sanjeev Rana, head of Korea Research at CLSA told Bloomberg TV.