On August 26, A-share trading concluded with Biwin Storage Technology Co.,Ltd. (688525.SH) closing at 223.35 yuan, down 1.96%, with a total market value of approximately 105.3 billion yuan. This storage chip company, valued at over 100 billion yuan, has seen its stock price retreat more than 56% from the historical high of 517 yuan reached in June. Behind this weak performance lies a classic "sell the news" scenario. Just after the market closed on August 24, the company released an impressive semi-annual report: revenue of 15.575 billion yuan, a year-on-year surge of 298.10%; net profit attributable to shareholders of 7.166 billion yuan, reversing from a loss of 226 million yuan in the same period last year. However, this stellar report card not only failed to boost the stock price but instead became a signal for major funds to exit—on the day of the earnings release, the stock fell 5.6% on heavy volume, with main capital seeing a net outflow of 745 million yuan in a single day. The stronger the earnings, the weaker the stock price—this abnormal divergence has brought the pricing dilemma of semiconductor cyclical stocks into the spotlight. Through interviews with industry experts and cross-referencing financial report data, we dissect the cyclical undercurrents and inventory chess game behind this storage "big bull stock."
Why can't 7.1 billion yuan in profit buy a single limit-up day? A net profit of 7.166 billion yuan, corresponding to a 3273% year-on-year increase, would be a phenomenal explosion in any industry. Biwin Storage attributes its high growth to "the AI computing power explosion and the storage industry entering a high-prosperity cycle," with AI emerging edge-side storage product revenue reaching approximately 2.86 billion yuan, a surge of 433.58% year-on-year. However, the capital market's reaction has been quite different. On the day of the earnings disclosure, the stock price plunged at the open, with main capital net outflow reaching 745 million yuan throughout the day, accounting for 14.85% of total turnover. Although it rose slightly by 0.99% the next day, it continued to fall 1.96% on August 26, closing at 223.35 yuan. The impressive earnings not only failed to ignite bullish sentiment but instead became a rallying call for profit-takers to flee en masse.
"This is precisely the most typical pricing mechanism for cyclical stocks," said Huang Lichong, an internationally seasoned investment banker and president of Huisheng International Capital, pointing out bluntly. "The income statement tells you how much money was made in the past six months, but the stock price trades on how much will be left in the coming year. The most beautiful profits often appear precisely when the market begins to worry about the cyclical peak." Industry data provides a precise footnote to this judgment. According to TrendForce statistics, although storage prices remained at high levels in the second quarter, the "acceleration" of price increases has fully slowed. In the first quarter of 2026, general-purpose DRAM prices rose 93%-98% quarter-on-quarter, narrowing to 58%-63% in the second quarter, with third-quarter expectations further shrinking to 13%-18%; NAND Flash also dropped sharply from 85%-90% in the first quarter to 10%-15% in the third quarter.
"The first derivative is still positive, but the second derivative has already deteriorated—the stock market always trades the second derivative first," Huang Lichong told reporters. Yuan Shuai, deputy director of the investment department at the China Urban Development Research Institute, interpreted this "reflexivity" from a market psychology perspective. "The pricing logic of the capital market is never fully anchored to current landed performance. When market participants generally believe that prosperity has reached a阶段性 high, capital will naturally choose to cash out in advance," Yuan Shuai explained. This advance pricing mechanism is a form of self-correction in strongly cyclical industries, avoiding the concentrated release of risks during the subsequent downturn phase.
It's worth noting that this "sell-the-news" phenomenon is not unique to A-shares. In July this year, global storage giants Micron, SK Hynix, and SanDisk all experienced similar trends of strong fundamentals but early stock price declines. This is no longer merely a battle between bulls and bears on a single stock, but a collective prediction by the entire storage industry's capital on the cyclical turning point.
Can 18.1 billion yuan in inventory lock in profits? If the 7.166 billion yuan profit is Biwin Storage's "letter of allegiance" to the market, then the 18.168 billion yuan in inventory on the balance sheet is the Sword of Damocles hanging over investors' heads. As of June 30, 2026, the company's inventory book value surged 130.89% compared to 7.868 billion yuan at the end of 2025, piling up 10.3 billion yuan in inventory in just six months. Among this, raw materials reached 8.755 billion yuan, and finished goods reached 5.537 billion yuan. Accompanying the surge in inventory is a sharply deteriorating cash flow—net operating cash flow in the first half was -6.261 billion yuan, short-term borrowings climbed to 6.5 billion yuan, and long-term borrowings reached 8.6 billion yuan.
"This issue is far more important than the 7.1 billion yuan profit," Huang Lichong said in an interview. "During an upward cycle, inventory is a profit amplifier; once prices turn, the same inventory becomes a loss amplifier." He calculated for reporters: if the 18.168 billion yuan in inventory faces a 10% price pressure, it corresponds to approximately 1.8 billion yuan in losses; if it falls 20%, that's a 3.6 billion yuan hole. Compared to the 7.1 billion yuan semi-annual profit, this is a risk exposure sufficient to swallow all earnings. The company itself candidly stated in the "Risk Factors" section of its semi-annual report: if the storage wafer market experiences a significant downturn, it does not rule out further provisions for inventory write-downs, which would impact overall performance.
Facing market concerns, Biwin Storage's choice is an aggressive "long-term agreement lock-in." In March and June this year, the company signed long-term procurement agreements with two storage wafer manufacturers, with cumulative committed procurement amounts reaching 3.3608 billion US dollars, explicitly "locking quantity and price" with commitment periods lasting up to 24 months. The company claims this move aims to "smooth industry cyclical fluctuations and stabilize profit levels." But in Yuan Shuai's view, this strategic positioning harbors hidden dangers. "Long-term agreements are a double-edged sword. If industry prosperity shifts, high inventory faces impairment pressure while long-term agreements may also form rigid cost burdens." He emphasized to reporters that the key to judging the value of this strategy lies not in the level of inventory scale, but in whether the company can convert locked-up upstream resources into differentiated products for high-value-added scenarios, "rather than passively enduring in the general-purpose market."
Huang Lichong further pointed out concerns in accounting treatment. The company uses the moving weighted average method for inventory. As procurement prices rise quarter by quarter, new purchases continuously push up the average cost, meaning the cost advantage of early low-cost inventory "will not exist permanently." And of this 18.1 billion yuan in inventory, how much is already covered by orders, and how much uses fixed selling prices? The semi-annual report does not provide clear answers.
Can the "world's largest" label tame the strong cycle? Can AI truly reshape storage from a "fate-determined strong cycle" into a "steadily rising growth stock"? This is the ultimate question regarding the rationality of Biwin Storage's 100-billion-yuan market value. The company's external narrative is filled with the label of "the world's largest AI emerging edge-side semiconductor storage solution provider." Its products are embedded in AI glasses and smartwatches from Meta, Google, Xiaomi, and Xiaotiancai, with AI edge-side revenue surging 433%—these facts do provide footnotes to the growth logic. But industry experts are far more sober.
Huang Lichong maintains the cautious stance typical of an investment banker toward the "world's largest" label: "I haven't seen sufficiently transparent, independent third-party global market share data to support this conclusion. I would split the story into 'real industrial opportunities' and 'capital market labels.'" He warned that even if AI raises the demand baseline, "AI will not cancel the economic cycle." The iron law of the semiconductor industry has never changed: high profits create supply, high valuations create capital expenditure, and supply ultimately recreates the cycle.
Yuan Shuai provided a more dialectical industrial perspective. He believes that the "sustained structural increment" brought by AI edge-side demand is indeed breaking the purely cyclical nature that previously relied solely on consumer electronics. "The proliferation of AI glasses and smart wearables is not a short-term theme speculation but a long-term trend of technological penetration." However, he also acknowledged that hard-tech companies must find a dynamic balance between short-term performance delivery and long-term value accumulation in commercialization and technological deepening.
In this debate between the "cyclical school" and the "growth school," the market has already voted with its feet. Although the year-to-date gain remains nearly doubled, the 56% pullback, a trailing price-to-earnings ratio of only 12.78 times, and a massive 9.5 billion yuan in margin financing balance all reveal the market's deep apprehension about the "cyclical peak." As of the close on August 26, Biwin Storage's market value stood at 105.3 billion yuan. On August 11, the company announced a share buyback plan of 200 million to 250 million yuan, with a maximum buyback price as high as 468.24 yuan per share, more than 110% above the current price, in an attempt to demonstrate long-term confidence. But until the execution pace of the 3.36 billion US dollar long-term agreements, the digestion progress of the 18.1 billion yuan inventory, and the turning point of the storage price "second derivative" turning from positive to negative become clear, this drama of divergence between earnings and stock price is destined not to end easily.
"For cyclical stocks, the lowest P/E ratio is often not when they are cheapest, but rather when earnings are at the cyclical peak," Huang Lichong said in the interview—perhaps the most sober footnote to Biwin Storage's current predicament.
Comments