Market Flash: AI Stocks Plunge Again — What Is the Market Really Worried About?

Last week, global semiconductor stocks faced intense selling pressure, with the Philadelphia Semiconductor Index plunging approximately 11% in a single week. Asian semiconductor markets, including Taiwan and Japan, followed suit with consecutive declines. Notably, most recently reported financial results from technology and semiconductor firms have exceeded expectations, yet stock prices continue to weaken. What exactly is the market worried about?

1. South Korea Tightens Leverage ETF Regulations, Sparking Semiconductor Stock Sell-Off

One of the earliest triggers of this global semiconductor correction came from South Korea. The country’s regulatory authorities admitted that previous approvals for leveraged ETFs tied to individual tech stocks like Samsung Electronics and SK Hynix were too hasty. They subsequently announced a suspension of new listings for certain single-stock leveraged ETFs and raised the minimum margin requirements for retail investors trading these products, aiming to reduce market volatility. Prior to this, Korean retail investors had heavily leveraged semiconductor stocks through margin financing and 2x leveraged products. After the regulatory tightening, concerns arose that investors would be forced to deleverage and liquidate positions, leading to concentrated selling of major stocks like Samsung Electronics and SK Hynix. This caused the Korean stock market to fall sharply first, accelerating existing profit-taking and deleveraging pressures, and rapidly spreading the global sell-off in AI and semiconductor stocks.

2. Fundamentals Remain Strong — The Market Is Correcting High Valuations

It's important to note that this stock price decline is not due to disappearing AI demand, but rather occurs despite solid corporate fundamentals. TSMC’s latest Q2 financial report showed a quarterly net profit of NT$706.6 billion (approximately USD 21.99 billion), up 77% year-on-year — not only a record high but also exceeding the market’s forecast of NT$632.6 billion. Q2 revenue reached USD 40.2 billion, at the upper end of guidance and above the expected USD 39.94 billion. This marks the fifth consecutive quarter of record profits. TSMC also raised its full-year revenue outlook from “over 30%” to “slightly above 40%” and increased its annual capital expenditure from USD 52–56 billion to USD 60–64 billion — a roughly 14.8% increase — indicating that AI chip demand remains strong. However, after AI-related stocks surged over the past year, the market now believes most positive news is already priced in. Even strong earnings are no longer enough to push prices higher, making this a prime opportunity for some investors to lock in profits.

3. Is Selling First and Buying Back Later Really More Profitable? A Nearly 100-Year Backtest Reveals the Answer

Many investors wonder: “Should I sell first, wait for a deeper drop, and buy back when the trend improves — wouldn’t that earn more?” This strategy may seem to avoid further downside, but the real challenge lies in whether you can buy back in time and avoid missing the market’s rebound. To answer this, Chin Hung Buy Fund conducted a backtest using the S&P 500 Total Return Index (including dividends) from December 1927 to July 2026, comparing two strategies: “buy and hold” versus “sell after a 10% drawdown from peak, buy back when index reclaims the 60-day moving average.”

The results showed that, using a daily rolling window across different holding periods, the average cumulative return of long-term holding consistently outperformed short-term trading. Over a 30-year holding period, long-term holding achieved an average cumulative return of 2,241%, significantly higher than the 1,707% from short-term trading — a difference of 534 percentage points. This indicates that even if short-term strategies avoid some corrections, repeatedly exiting the market may cause investors to miss the crucial early rebound phase, resulting in lower overall returns. Rather than repeatedly guessing entry and exit points, maintaining long-term exposure during market volatility offers a better chance to benefit from compounding.

Source: Bloomberg, Chin Hung Buy Fund analysis. Daily data from S&P 500 Total Return Index, December 1927 to July 2026. Short-term strategy: sell after 10% peak-to-trough drop, buy back when index crosses above 60-day moving average. Daily rolling backtest, showing average cumulative returns over 10-, 20-, and 30-year holding periods. This data is a historical simulation only and does not guarantee future investment performance. Results may vary under different index movements, allocations, and timeframes. Investors entering at different times will experience different performance outcomes, and past performance does not guarantee future results.

Chin Hung Investment Strategy

AI Long-Term Trend Unchanged — Correction Tests Investment Discipline

Based on recently reported earnings and AI capital expenditures, the long-term growth trend of AI remains intact. This correction appears to be a normal valuation adjustment rather than a fundamental reversal. Instead of trying to time the bottom, investors should establish long-term discipline, using dollar-cost averaging or the “Super Bottom King” auto-bottom-fishing mechanism to steadily accumulate positions during market volatility.

Chin Hung Featured Funds

A23042 Fuh Hwa US New Star Fund - TWD Subscribe

B16008 Invesco Global Entrepreneur Fund A USD Subscribe

Future Tech King Subscribe

“Chin Hung Buy Fund” Operates Independently

This information is for reference only. Chin Hung Buy Fund has made every effort to provide accurate opinions and information from reliable sources, but cannot guarantee the completeness of such data.

Under the latest Financial Consumer Protection Act regulations, to provide better investment services, our company implements an “Investor Attribute Assessment” procedure. By completing the “Investment Risk Suitability” assessment on the Chin Hung Buy Fund website, we will apply product suitability rules based on your profile and request you to update it at least annually. Failure to complete the assessment may affect your trading ability.

Fund investments are intended for long-term purposes and should not expect high short-term returns. Fund prices and returns may fluctuate, and investors may not recover their full investment amount. Investors should understand and assess independently before subscribing.

For funds denominated in foreign currencies, investors bear the exchange rate risk when converting dividend distributions or redemption proceeds back to New Taiwan Dollars. If the NTD exchange rate strengthens against the foreign currency at conversion compared to the original investment date, investors may incur exchange losses.

Each fund employs mechanisms such as “Fair Value Adjustment” (or “Anti-Dilution”), short-term trading rules, and “Price Adjustment Mechanism” (or “Swing Pricing”). For details, please refer to each fund’s prospectus. Investors must comply with the relevant trading rules and restrictions outlined in the fund prospectus.

Each fund has been approved or recognized by the Financial Supervisory Commission, but this does not imply zero risk. The fund manager’s past performance does not guarantee minimum investment returns; the manager owes a duty of care but is not liable for fund profits or losses, nor does it guarantee minimum returns. Investors should carefully read the investor handbook and fund prospectus before subscribing. Investors should understand and assess independently.

Fund-related expenses (including distribution fees) are disclosed in the fund prospectus and investor handbook. Investors can access this information via the Public Information Observation Station or Fund Information Observation Station.

Due to their non-investment-grade credit ratings (or lack thereof) and high sensitivity to interest rate changes, non-investment-grade bond funds may suffer losses from rising interest rates, reduced market liquidity, or issuer default, non-payment of principal/interest, or bankruptcy. These funds are unsuitable for investors unable to bear such risks. Investors should not allocate an excessive portion of their portfolio to funds targeting non-investment-grade bonds. Some funds or their underlying bond sub-funds may invest in U.S. Rule 144A bonds, which are private placements and carry higher risks of illiquidity, incomplete financial disclosure, or price volatility due to opacity. Investors should be aware of these risks.

Funds invest in loss-absorbing bonds issued by financial institutions (including Contingent Convertible Bonds and Total Loss-Absorbing Capacity bonds, or CoCo and TLAC bonds). If a financial institution’s capital adequacy falls below a certain level or faces severe operational or bankruptcy crises, these bonds may be written down or converted into equity by contractual or statutory mechanisms, potentially resulting in partial or total loss of principal, cancellation of interest, conversion to equity, changes to maturity, coupon, payment dates, or suspension of distributions.

Fund distributions do not represent actual fund returns, and past distributions do not guarantee future ones. Fund net asset values may fluctuate due to market factors.

The fund distribution yield is calculated as: (Distribution per unit ÷ NAV on the day before ex-dividend date) × 100%. The total return including dividends is calculated as: [(NAV at month-end of ex-dividend month + Distribution per unit) ÷ NAV at end of previous month / 3 months / 6 months / 1 year - 1] × 100%.

Some fund distributions are paid before deducting related expenses. Fund distributions may be paid from fund earnings or principal. Any distribution sourced from principal may erode the original investment amount. Information on whether a fund pays distributions from principal is available on the websites of respective asset management companies and master agents.

Funds investing in emerging markets may face higher volatility and risk levels, along with potential political and economic risks.