(Bloomberg) -- SK Hynix Inc. is accelerating the launch of its next-generation AI memory chips after Nvidia Corp. Chief Executive Officer Jensen Huang asked the South Korean supplier if it can provide samples six months earlier than originally planned, signaling persistent global demand for cutting-edge semiconductors. 

Huang raised the request regarding the upcoming high-bandwidth memory chips, or HBM4, during a recent meeting with SK Group Chairman Chey Tae-won, the South Korean tycoon told reporters on the sideline of the SK AI Summit in Seoul on Monday. Chey said Huang has succeeded as a leader in part because of his emphasis on speed. 

Consequently, SK Hynix said earlier this month that it is on track to supply customers with HBM4 in the second half of 2025, reaffirming a time line first offered in a post-earnings call in October. The South Korean company is the primary supplier of the HBM chips Nvidia depends on to make the US company’s artificial intelligence chips work. 

Both Nvidia and SK Hynix have benefited from the insatiable appetite major companies and governments have for the Santa Clara, California-based company’s cutting-edge chips, the gold standard for training AI algorithms. 

Whether SK Hynix can indeed deliver the next-generation AI memory in 2025, however, hinges on whether the South Korean chipmaker can complete Nvidia’s complex qualification process in time. “Both sides are on the same page” on the HBM4 schedule, Chey told reporters. 

Meanwhile, Nvidia still cannot make enough AI chips to fulfill demand, Chey said during the company event, and SK Hynix is working with the US company to resolve the ongoing supply shortage.

SK Hynix also disclosed its roadmap for some additional new products on Monday. The chipmaker is planning to introduce 16-layer HBM3E in early 2025 and release HBM5 and HBM5E between 2028 and 2030, the chipmaker’s Chief Executive Officer Kwak Noh-Jung said during the AI summit. SK Hynix closed up 6.5% on Monday. 

SK Hynix last month reported a record quarterly operating profit and said it plans to begin supplying its top-of-the-line 12-layer HBM3E in the fourth quarter. 

Samsung Electronics Co., which has struggled to catch up with SK Hynix on the AI chip front, has said it plans to mass-produce HBM4 chips in the second half of next year. Samsung is hoping its HBM4 products will help the company be competitive against SK Hynix in supplying Nvidia. 

©2024 Bloomberg L.P.