SK Group Chairman Pushes Back on Nvidia Dependence Concerns as SK Hynix Stock Jumps

SK Group Chairman Chey Tae-won on Thursday sought to calm concerns that SK Hynix’s revenue and outlook are too dependent on Nvidia Corp., arguing the chipmaker remains part of a wider AI ecosystem rather than a single point of risk. The remarks, carried by market coverage, came as investors reacted to the idea that any Nvidia-related slowdown could quickly ripple through high-bandwidth memory and other DRAM and NAND product markets that are tightly linked to AI servers.

Chey’s comments were reported in the context of heightened attention on customer concentration in semiconductors, especially among suppliers of memory components that are used to build data-center systems. In the reporting, the chairman dismissed the premise behind the worry: that SK Hynix’s exposure to Nvidia is the central driver of its business trajectory.

The market reaction was swift. SK Hynix shares rose sharply on the news, with the coverage framing the move as roughly an 8% gain. The jump reflected how quickly equity investors can shift from skepticism to confidence when leadership directly addresses a risk they had been focused on.

The framing in the market report emphasized the question investors were effectively asking: if Nvidia demand is the key engine behind AI server buildouts, then does a memory supplier that benefits disproportionately become vulnerable to Nvidia’s own purchasing behavior. Chey’s message, as described, was that this line of thinking is too narrow, implying that SK Hynix’s position is supported by the broader AI and data-center supply chain rather than being tied to one company’s fate.

From a sector standpoint, memory makers sit at a crossroads of two forces. One is the pace of AI data-center spending, which can accelerate demand for higher-performance memory used for training and inference. The other is the inherently cyclical nature of memory pricing, where supply expansions, inventory levels, and contract dynamics can swing results even if AI demand remains strong.

Nvidia’s role in AI infrastructure is central, but the debate for investors is how that centrality translates into customer concentration risk for upstream and adjacent suppliers. When a large customer is widely viewed as a dominant end-market catalyst, equity markets often apply a “single-customer” lens, even if the supplier sells into a larger network of server original equipment manufacturers, system integrators, and module builders.

Chey’s reported dismissal of the dependence concern suggests SK Group and SK Hynix are trying to manage that narrative risk as much as the underlying operational risk. However, the market coverage also indicated limits to what investors were given in the form of numbers, contract detail, or forward guidance. The reporting did not provide a breakdown of revenue by customer or a quantified hedging or diversification strategy, at least not in the portion of information referenced in the market item.

Going forward, investors are likely to keep pressing for clarity on exposure: whether SK Hynix’s AI-related demand is spread across multiple buyers and platforms, and how management thinks about pricing and supply planning when AI spending remains uneven over time. The next set of disclosed financial results, guidance, or management commentary on demand visibility and customer mix would be the most direct place to test whether the “not worried” message can be supported by measurable data.