Performance clearly hasn’t collapsed, yet tech stocks have plunged sharply? The harshest truth of China’s A-share market: It’s not that companies are underperforming—it’s that there are too many investors and prices are too high.

Performance clearly hasn’t collapsed, yet tech stocks have plunged sharply? The harshest truth of China’s A-share market: It’s not that companies are underperforming—it’s that there are too many investors and prices are too high. Some people call this ‘hindsight bias’—and I despise hindsight bias most of all. I never engage in it. On June 30, 2026, I published an article warning that tech stocks were overvalued—see the screenshot below. On August 1, 2026, I again commented on high-flying tech stocks. Recently, many investors have been left utterly bewildered by market movements. Open up the latest earnings reports: Companies in semiconductors, AI hardware, and the STAR Market sector—not only haven’t seen performance collapse, but many are posting exceptionally strong growth rates. Revenue is steadily rising; profits are being delivered; industry prosperity remains intact—there is virtually no substantive fundamental headwind. Yet as everyone has witnessed: Indices continue to retreat; individual stocks keep plunging; many investors holding companies with realized profits are still suffering losses amid the broad selloff. So everyone is asking the same question: If companies are generating solid profits, why is the stock price falling so sharply? The answer is just two words: Too expensive. This round of tech-stock correction has nothing to do with corporate operational health. At its core, it’s not a collapse in fundamentals—it’s a valuation reset; not a company problem—it’s an extreme overextension of market expectations. I. Absurd valuations: You’re not buying profits—you’re buying stories about the next few years. At the peak of this rally in early July, the level of bubble formation in the tech sector ranked among the highest ever seen in A-share history. Here are several hard data points (source: Tonghuashun iFinD, as of market close on July 3, 2026): - The STAR 50 Index’s PE (TTM) surged to 231x—placing it at the 99th percentile of its historical range since inception; - The semiconductor sector’s weighted P/E ratio peaked above 200x, reaching an all-time extreme in historical terms. Many retail investors misunderstand valuation: They assume high valuation = high growth potential, superior sector quality, and long-term holding worthiness. In reality, the true meaning of high valuation is brutally simple: Today’s stock price has already fully priced in—and front-loaded—all optimistic expectations for the next three, five, or even more years. Investors who heavily bought into STAR Market, chip, and AI computing stocks in early July weren’t buying 2026’s actual earnings—they were buying a flawlessly scripted capital-market story. That story’s plot is straightforward: AI demand will explode forever; compute capacity will always be in short supply; tech hardware will sustain ultra-high growth indefinitely; the sector faces no ceiling. Amid extreme emotional consensus, the market assumed this story would hold eternally—so valuations kept climbing, pushing asset premiums to absurdly extreme levels. But one immutable law of equity markets spanning centuries holds true: When the story is most perfect—and everyone believes it unquestioningly—that’s precisely when stock-price risk peaks. II. The backlash of crowded trading: No new money is the biggest bearish catalyst. Stock investing often isn’t just about corporate value—it’s also about shareholding structure and market liquidity. Even the most outstanding growth assets inevitably face valuation corrections once capital becomes excessively concentrated and trading becomes extremely crowded. From late June to early July, A-share market crowding reached a multi-year extreme: - The daily trading volume share of the entire TMT sector peaked near 49.2%, meaning nearly half of all market turnover was funneled exclusively into AI, semiconductors, and computing infrastructure; - The top ten core tech leaders by trading volume collectively accounted for over 50% of the market’s total daily turnover.

Why does the stock price plunge after earnings are released?

On August 5, after U.S. market hours, two major memory chip giants—SanDisk and Western Digital—released their quarterly earnings simultaneously: ・ SanDisk: Q4 revenue of $8.96 billion, up 372% year-on-year—far exceeding market expectations; approved a $14 billion share buyback program. ・ Western Digital: Revenue of $3.747 billion, up 44% year-on-year; net profit surged 12-fold—also significantly beating expectations. Yet the result? SanDisk fell 7% after hours; Western Digital plunged 11%. This isn’t the first time. In July, SK Hynix reported its Q2 operating profit surged 18-fold year-on-year—and plummeted 30% intraday. The better the earnings, the steeper the drop. This is the ‘magical realism’ of the memory chip sector. Why? Because memory chips are a classic cyclical industry. Memory chip prices swing like a rollercoaster through distinct cycles: Prices surge and companies reap windfall profits during supply shortages; prices crash and firms post losses during oversupply. A full cycle typically lasts 3–4 years. Cyclical stocks follow a pricing logic diametrically opposite to that of typical growth stocks: Stage Industry Condition Stock Price Behavior Cycle Bottom Industry-wide losses, widespread despair Price quietly bottoms out and begins rising Early Upswing Earnings shift from loss to profit Price accelerates upward—largest gains occur here Mid-Upswing Earnings sustain high growth, consistently beat expectations Price continues rising—but at a decelerating pace Peak Prosperity Earnings hit record highs, but growth rate slows Price tops out and begins declining Early Downturn Earnings still rise—but growth momentum weakens Price falls sharply Late Downturn Industry-wide losses return Price hits bottom—awaiting next cycle The core rule is just one sentence: Stock prices always lead earnings by 6–12 months—so the best earnings results often coincide precisely with the stock price peak. Why? Because savvy investors quietly take profits and exit precisely when ‘everyone is bullish and earnings look brightest’—they know that after the cyclical peak comes the downturn. Where are SanDisk’s ‘cracks’? Specifically for this memory cycle, the market is now concerned about: 1. Slowing price-increase momentum: DRAM contract prices rose over 90% quarter-on-quarter in Q1, but are expected to rise only 13%–18% in Q3—the inflection point in growth rate is already evident. 2. Weak consumer demand: While AI server demand remains robust, consumer electronics—including smartphones and PCs—have yet to see broad-based recovery. A cycle propped up by a single pillar raises sustainability concerns. 3. Supply-side risks: Samsung and SK Hynix, emboldened by high profitability, are showing stronger willingness to expand capacity—potentially increasing future supply. 4. Guidance misses: SanDisk’s revenue guidance for next quarter—midpoint $10.55 billion—fell short of the market’s $11.16 billion expectation—the final straw that broke the camel’s back. Once the market senses ‘the most profitable phase may already be behind us,’ the stock price drops immediately—even if earnings are still rising. II. Four Core Reasons: Why ‘Good Earnings, Falling Stock Price’ Is So Common Memory chips represent the most extreme case—but ‘good earnings, falling stock price’ is by no means unique to memory. It has occurred across technology, consumer staples, pharmaceuticals, new energy, and nearly every sector. Four underlying reasons explain this phenomenon:

Hudson Technologies Q3 2026 Earnings Miss Sends Shares Lower

Oct 28, 2026

Hudson Technologies missed Q3 2026 earnings and revenue targets, causing a 10% share drop amid lower HFC prices. The firm plans to focus on profitable sales and new separation technology investments.

Investing.com Canada

2026-05-06

Hudson Technologies Reports First Quarter 2026 Results

2026-05-06

Q1 2026: Hudson Technologies (NASDAQ: HDSN) grows revenue 9%, EPS falls

2026-05-06

Hudson Technologies' Q1 CY2026 Sales Beat Estimates

2026-05-06

Hudson Technologies Q1 2026 Earnings Miss EPS Forecast

2026-05-07

Warm Western weather lifted Hudson sales, but Q1 profit fell