Nam Hyun-woo has worked as a staff writer at The Korea Times since 2013, mostly covering business and politics. He currently belongs to the Business Desk where he covers topics such as emerging tech, AI, ICT and Korea's chaebol community. Prior to joining the team, he was the paper's correspondent for the presidential office of Korea during the Yoon Suk Yeol and Moon Jae-in administrations.
Nvidia’s upcoming earnings to sway Korean chipmakers’ strategy
Nvidia CEO Jensen Huang, center, speaks at the GeForce Gamer Festival in Seoul’s Gangnam District, Oct. 30. From left are Hyundai Motor Group Executive Chair Chung Euisun, Huang and Samsung Electronics Executive Chairman Lee Jae-yong. Yonhap
By Nam Hyun-woo
AI outlook to determine Samsung, SK hynix fab strategies amid global investment frenzy
The global tech industry is paying keen attention to Nvidia’s upcoming earnings report, as it is expected to be the bellwether for Samsung Electronics, SK hynix and other domestic semiconductor firms’ strategies for expanding their capacity.
Industry officials say the announcement will be crucial to verifying whether the current investment frenzy over artificial intelligence (AI) is overhyped. Depending on the announcement, the current cash flow between Nvidia and hyperscale data center operators could be affected, potentially leading chipmakers to review the pace of their fab expansion plans.
U.S.-based Nvidia is set to announce its 2026 third quarter earnings on Wednesday (local time) after the close of markets. Analysts are anticipating $55.2 billion in revenue, up 57 percent year-on-year, with an adjusted earnings per share of $1.26, up 55 percent from a year earlier.
The growth rate is considered a key indicator of the genuine global market demand for AI. In recent months, Nvidia has faced questions over whether expectations have outpaced actual demand and whether the growing scale of AI investment is translating into real profit.
If the company continues to show sharp growth in its earnings, confidence for AI will remain high. But if the pace slows, it could reinforce concerns that the market has been overly optimistic. Nvidia has posted revenue growth for six consecutive quarters, but its growth rate has gradually slowed.
“Expectations are steep, and the reaction will depend on whether the company proves the AI boom is still delivering expanding profitability rather than just expanding scale,” said Nigel Green, CEO of financial advisory organization deVere Group.
Green added that investors are paying close attention to details such as data center contribution and the flow of orders from hyperscalers. A small miss in any of those areas “could change sentiment instantly, especially after one of the strongest multiyear share price performances in modern market history.… When expectations tower over reality, even a strong quarter can fall short of what the market has imagined.”
Some doubt has already materialized. Thiel Macro, a hedge fund led by PayPal and Palantir Technologies co-founder Peter Thiel, offloaded its entire stake in Nvidia on Monday, following SoftBank’s earlier unloading of its Nvidia stock, sounding alarms over a possible AI bubble.
If earnings growth falls short of expectations, it would suggest that actual demand for Nvidia’s AI accelerators is not matching market forecasts, which would affect the expansion plans of Samsung Electronics and SK hynix regarding high-bandwidth memory (HBM), an advanced memory chip used in Nvidia’s AI accelerators.
Samsung Electronics' semiconductor campus in Pyeongtaek, Gyeonggi Province / Courtesy of Samsung Electronics
Samsung Electronics and SK hynix are both scaling up their HBM capacity. Samsung said Sunday it decided to resume construction on its P5 line at its semiconductor campus in Pyeongtaek, Gyeonggi Province, with the goal of beginning operations in 2028. SK hynix is expediting the operation of its new M15X fab in Cheongju, North Chungcheong Province, as well as building a megasize chip cluster in Yongin, Gyeonggi Province, which is expected to complete its first fab in 2027.
These investment decisions are based on the current AI demand, but the new fabs won’t be fully operational until two to three years later. If questions arise about future demand for Nvidia’s accelerators, strategies for these new fabs could be reconsidered.
Investors are also awaiting more insights from Nvidia CEO Jensen Huang regarding the sales outlook for Nvidia products based on its next-generation architecture, Rubin.
Nvidia said in August that it had taped out its Rubin graphics processing units (GPUs) and sent them to TSMC for production. Mass production is expected to begin around the third quarter of next year, meaning the schedule will coincide with the full-fledged operation of Samsung Electronics and SK hynix’s new fabs. This means the new fabs’ utilization rate could be swayed by the availability of Rubin-based GPUs.
Despite lingering doubts, domestic analysts expect the current semiconductor upcycle to maintain its momentum. “Nothing has changed in the memory market,” said Samsung Securities analyst Lee Jong-wook. “The strong data center demand next year, supply shortages, longer contract periods and sharp growth in prices all remain intact as investment drivers.”