[Edaily Reporter Kim Hyung-il ] Expectations are growing that the artificial intelligence (AI) investment cycle will be prolonged, as NVIDIA has released medium- to long-term growth forecasts that significantly exceed market expectations. With the expansion of AI data centers expected to drive increased demand for memory—including High Bandwidth Memory (HBM)—attention is focused on whether SamsungElectronics(005930)and SK hynix(000660)will also benefit.

According to the securities industry on the 29th, NVIDIA’s stock price closed at $227.98 on the New York Stock Exchange on the 27th (local time), up 8.74% from the previous trading day. Its market capitalization reached approximately $5.52 trillion, an increase of $442 billion in a single day. This marks the second-largest single-day increase in market capitalization for an individual stock in history.

The rise in the stock price was driven by both earnings and the outlook for medium- to long-term growth. NVIDIA projected that revenue for fiscal year 2028 would increase by approximately 70 percent—significantly exceeding market expectations of about 44–45 percent. NVIDIA explained that while demand for AI infrastructure is growing rapidly, it is unable to fully meet that demand due to supply constraints, such as those related to memory.

Earnings were also robust. NVIDIA’s revenue for the second quarter of fiscal year 2027 was $96.221 billion, a 105.9% increase year-over-year. Operating income was $63.956 billion, and earnings per share (EPS) was $2.22. However, operating income and EPS are on a non-GAAP basis; on a GAAP basis, operating income was $63.734 billion and EPS was $2.46. Data center revenue reached $89 billion, a 117% increase from the previous year. Third-quarter revenue guidance was set at $108 billion, exceeding market expectations.

This growth trend is seen as a positive sign for memory manufacturers as well. NVIDIA’s commitments related to supply and production capacity have expanded from the previous $119 billion to $279 billion. Since a significant portion of this is believed to be related to securing memory, including HBM, the likelihood of continued memory supply shortages is increasing.

Rising memory prices are weighing on NVIDIA’s profitability. NVIDIA projected a gross profit margin (GPM) of 74% for the third quarter and expects it to decline to 71–72% in the fourth quarter. Rising memory prices are cited as the primary factor contributing to this pressure. Conversely, from the perspective of memory manufacturers, rising prices and increased investment in AI infrastructure could serve as drivers of improved performance.

SK hynix stands to benefit directly from growing demand for HBM. MIRAE ASSET SECURITIES maintained its “Buy” rating and target price of 2.8 million won for SK hynix, forecasting that the memory supply-demand balance will remain tight through 2028. The firm anticipates that Chinese accelerator manufacturers will increasingly adopt HBM in their next-generation products, leading to rising HBM demand even in the AI accelerator market beyond NVIDIA.

SamsungElectronics is also expected to benefit from increased investment in AI data centers. KB Securities projected that AI data center capacity in the U.S. and China will more than double, rising from 95 GW in 2026 to 201 GW in 2030. The firm projected that, driven by the advancement of AI models and growing demand for inference, not only will data center capacity expand, but computational performance per server and memory capacity will also increase. In particular, due to differences in computational efficiency among Chinese AI accelerators—which may require more accelerators and memory to process the same computational load—the analysis concluded that China’s expanding AI investments will have a positive impact on demand for SamsungElectronics’ memory products.

In fact, NVIDIA’s moves to expand supply are already being reflected in the earnings outlook for domestic memory manufacturers. Based on NVIDIA’s expansion of supply chain commitments and rising demand for HBM, there is speculation that profit forecasts for SamsungElectronics and SK hynix may be revised upward. However, since NVIDIA’s increased AI investment does not necessarily translate directly into improved earnings for domestic companies, it is necessary to monitor both the expansion of HBM supply and memory price trends.

Ultimately, the key question is how quickly NVIDIA’s increased AI investments will translate into improved earnings for domestic memory manufacturers. Kang Da-hyun, an analyst at KB Securities, predicted that expanded investment in AI data centers will be accompanied by increases in computational performance and memory capacity per server. If AI infrastructure investment expands, led by NVIDIA, and HBM demand rises, expectations for improved earnings at SamsungElectronics and SK hynix are expected to grow even further.