Chip stocks are climbing once again! Today (February 19th), led by semiconductor shares, the South Korean stock market experienced a significant upswing. The KOSPI index surged by over 3% at one point, reaching a new all-time high. Samsung Electronics saw gains exceeding 5%, while SK Hynix rose over 2%.

Reports indicate that the price for Samsung Electronics' new-generation high-bandwidth memory chip, HBM4, has surged to $700, representing a 20% to 30% increase compared to HBM3E. SK Hynix, which is expected to begin mass production of HBM4 soon, is anticipated to set its price at a similar level. For context, in August of last year, the unit price for HBM4 supplied by SK Hynix to NVIDIA was approximately $500.

The strong performance of chip stocks propelled the South Korean market to a new peak. On Thursday, most Asia-Pacific markets advanced, with South Korean equities showing the most robust gains. At the time of reporting, the KOSPI index was up 2.68% to 5,653 points. Memory chip concept stocks led the gains, with Samsung Electronics rising 4.72% and SK Hynix increasing by 2.16%. Heavy industry and construction sectors also moved higher, with Hyundai Heavy Industries gaining over 7%, Doosan Group rising nearly 6%, and Daewoo E&C and Hyundai Engineering & Construction both climbing over 4%.

According to international media, amid ongoing supply shortages for memory chips, the reported supply price for Samsung's new HBM4 has risen to $700 (approximately 1.01 million Korean won). SK Hynix, poised to start HBM4 mass production, is expected to price its product similarly. This has led market forecasts to predict that both companies could achieve operating profits of around "30 trillion won" for the first quarter of this year.

It is reported that Samsung Electronics commenced mass production of HBM4 ahead of global competitors starting on the 12th of this month and is currently negotiating prices approximately 20% to 30% higher than the previous HBM3E generation, around $700. Just six months prior, in August, the price for HBM4 supplied by SK Hynix to NVIDIA was around $500, indicating a sharp price increase in a short period.

Market analysis suggests that as the profitability of general-purpose DRAM has improved significantly, now rivaling that of HBM, Samsung and SK Hynix have enhanced bargaining power and no longer need to focus exclusively on HBM production. In practice, Samsung Electronics is cautiously adjusting its output to avoid concentrating too much DRAM capacity on HBM, aiming to maximize overall profits.

With prices rising across HBM, general-purpose DRAM, and NAND flash memory, both Samsung Electronics and SK Hynix are projected to break their previous quarterly operating profit records in Q1 of this year. Samsung Electronics reported its first-ever quarterly operating profit exceeding 20 trillion won in Q4 of last year, while SK Hynix's profit for the same period was around 19 trillion won. Securities firms predict that Samsung's Q1 operating profit could reach 32 trillion won, with SK Hynix potentially exceeding 28 trillion won, officially ushering in a new era of "quarterly operating profits over 30 trillion won" for South Korean companies.

The earnings growth momentum for Samsung and SK Hynix is expected to continue throughout the year. Global investment bank Morgan Stanley forecasts that Samsung's full-year operating profit for this year will reach 245.7 trillion won, while SK Hynix is projected to achieve 179.4 trillion won, representing year-on-year increases of 464% and 280%, respectively.

In the US, technology stocks rebounded, with memory-related concepts showing strength. On Wednesday local time, as investor concerns about the disruptive impact of artificial intelligence eased, the three major US stock indices closed higher. The Nasdaq Composite rose 0.78%, the S&P 500 gained 0.56%, and the Dow Jones Industrial Average advanced 0.26%.

Memory concept stocks generally rose, with Micron Technology (MU) up 5.30%, Western Digital (WDC) gaining 4.38%, Seagate Technology rising 1.97%, and SanDisk increasing 1.66%. Reports noted that prominent hedge fund manager David Tepper increased his holdings in Micron Technology.

Major technology stocks also moved higher collectively. NVIDIA (NVDA) rose 1.63%, Amazon climbed 1.81%, while Microsoft and Meta Platforms (Facebook) gained over 0.60%. On the news front, Meta agreed to a new deal to deploy millions of NVIDIA chips in the coming years. Additionally, regulatory filings revealed that著名 investor Bill Ackman's Pershing Square fund increased its stake in Amazon by 65% during the fourth quarter of last year, making Amazon the fund's third-largest holding.

Ross Mayfield, an investment strategy analyst at Baird, commented, "At some point, weakness in tech stocks inevitably attracts marginal buyers. These are still high-growth assets; they were expensive before, and now they've become cheaper. Many will still want exposure to tech stocks in the coming years."

Following a sell-off in tech stocks triggered by concerns about AI's disruptive potential, investors are searching for signs of a market bottom. On Wednesday, a US chip stock index rose 1%, and an ETF tracking software companies jumped 1.3%.

Paul Stanley of Granite Bay Wealth Management suggested the sell-off in software stocks might be "overdone," characterizing it largely as a knee-jerk reaction as investors try to identify the winners and losers in the AI space. "While the potential of AI is vast, investors should not assume every company will succeed in the AI domain," Stanley stated.

In a recent report, Goldman Sachs mentioned that AI technology primarily functions as a powerful intelligence layer rather than a fundamental replacement. AI models still rely on Systems of Record to maximize their value. Particularly in enterprise settings, AI models require vast amounts of high-quality, structured, and historically accurate data for training and operation. Existing Systems of Record serve as repositories for this data. Established SoR companies like SAP, Salesforce, Oracle, and Workday have spent decades building robust data validation, governance, and compliance processes. For instance, AI applications for financial planning and analysis must draw accurate historical data from meticulously maintained and audited trusted ledgers.