NVIDIA CEO Huang: AI Chip Supply Constraints Persist; SK Chairman Vows to Double Memory Output

NVIDIA CEO Jensen Huang acknowledged ongoing supply limitations for artificial intelligence chips, signaling continued strong demand in the burgeoning AI market. Speaking at a global media conference in Taipei on Tuesday, Huang indicated that while NVIDIA is securing supply for "very solid growth," constraints remain.

“

We are securing supply for very solid growth, but there are still supply constraints.

This comes as NVIDIA expands its AI chip strategy beyond large data centers to consumer products, including laptops and desktops, with the launch of its new "Blackwell" platform. Despite these efforts, Huang believes the market is still experiencing a demand exceeding supply.

SK Group Chairman Choi Tae-won echoed concerns about future memory shortages, projecting they could persist until 2030 due to AI's increasing demand for memory. He announced plans to double the group's overall memory production capacity within the next five years, emphasizing readiness to invest in necessary facilities.

“

As AI continues to expand, more memory is needed.

Meanwhile, Samsung Electronics and SK Hynix, both key suppliers of High Bandwidth Memory (HBM) for NVIDIA's next-generation AI chips, engaged in a technological showcase at the Computex trade fair. Samsung unveiled a prototype of its HBM5, two generations ahead of current production, signaling its intent to lead the market. Choi expressed SK's desire to maintain its position as a primary NVIDIA supplier.

“

We will double the entire memory production capacity within the next five years.

Huang also expressed interest in expanding cooperation with South Korean companies in "physical AI," particularly in robotics, during a meeting with Korean business leaders. He plans to visit South Korea on June 5th and is scheduled to appear on the TV show "You Quiz on the Block."

“

We have been a major supplier to NVIDIA and hope to remain so.