半導体・先端技術研究｜Semiconductor Research | TokyoBrainie

English · 日本語

English｜A sell-off at the open, a recovery into the close

My provisional interpretation is that rebound buying after a sharp decline may have been supported by resilient analyst expectations and reports of stronger SSD prices. The dominant cause remains unverified. The comparison also changes the original question: Kioxia was not the only semiconductor-related stock to recover substantially after opening lower.

1. Two returns describe two different experiences

Kioxia opened at JPY 12,705 at 09:06, against a previous close of JPY 13,375. Its JPY 12,700 daily low was recorded in the same minute. It closed at JPY 13,600: a 7.04% gain from its own opening price and a 1.68% gain from the previous close. These are different measurements. [P1]

The comparison table shows that Advantest also gained 6.23% from its opening price, despite finishing 0.23% below its previous close. Its opening gap was deeper. All six stocks finished above their respective opening prices. Kioxia stood out as the only positive close-to-close return within this limited sample, rather than as the only stock with a substantial recovery. \[P1,P2\] This is a retrospective six-stock comparison, not an industry-wide ranking, a sector index or a record of selections fixed before trading. Each opening price is the stock’s actual first trade; openings need not occur simultaneously. A synchronized relative-strength test requires common timestamps and intraday data. ### 2. A timeline prevents a story from becoming a cause Kioxia had already fallen approximately 23.59% between its 9 October close of JPY 17,505 and its 16 October close of JPY 13,375. That creates a plausible setting for a rebound; it does not establish undervaluation. \[P1\] Before the 19 October open, Fisco’s 08:25 technical commentary discussed the 52-week moving average and possible oversold conditions. This documents an ex-ante rebound argument, but does not validate that argument as a trading edge. At 09:05, Traders Web reported selling pressure associated with US technology weakness and CoreWeave’s decline. \[N1,N2\] At 11:05, IFIS reported a maintained bullish rating. Fisco’s 11:21 article identified Goldman Sachs and a target-price revision from JPY 38,667 to JPY 39,000. The underlying research report has not been obtained. Publication times are not proof of when clients first received information, or when buying began. \[N3\] The SSD pricing story was already in circulation: Fisco’s 18 October roundup referred to a 13 October topic and cited Nikkei reporting that large-lot PC SSD prices had risen approximately 20% quarter on quarter in July–September. The original pricing data has not been checked. This is background support, not a newly discovered 19 October catalyst. \[N4\] ### 3. Three hypotheses, with different evidentiary limits A rebound after heavy selling. Traders can seek a short-term recovery without concluding that long-term risks have disappeared. If selling is concentrated at the opening auction and subsequent supply eases, prices can recover. The daily bar is compatible with this mechanism, but cannot prove that forced selling ended or identify the buyers. Business expectations remained resilient. Bullish analyst coverage and SSD price reports may have supported confidence. However, the target-price increase was only about 0.86%. A high absolute target is not equivalent to a large amount of new information. Nor can a reported PC SSD price increase be copied directly into Kioxia’s company-wide selling price or earnings forecast. \[N3,N4\] A shared recovery with company-specific differences. Every stock in the sample closed above its opening price. A common recovery component therefore deserves consideration alongside individual catalysts. Short covering is an alternative explanation, but there is no verified split between new long positions and purchases closing shorts. \[P1,P2\] ### 4. The semiconductor label hides different earnings mechanisms Kioxia’s official product materials describe NAND flash and applications including enterprise storage, data centers, PCs and mobile devices. That establishes its product exposure; it does not explain today’s share-price move. \[O1\] For a NAND producer, I would examine realized prices, bit shipments, product mix, manufacturing costs, investment and cash generation. For equipment suppliers, customer budgets, orders, deliveries and revenue-recognition timing enter through a different route. Shared AI demand does not imply identical earnings sensitivity. Today’s divergence is a reason to investigate those differences, not proof that they caused it. ### 5. Volume and VWAP require careful interpretation Kioxia traded 56,032,900 shares, approximately 0.97 times the previous session’s 57,744,400 shares. Its large turnover should not be described as a day-on-day volume surge. No 20-day volume baseline or RVOL test has been calculated. \[P1\] The recorded final VWAP was JPY 13,352.46, putting the close approximately 1.85% above that session’s volume-weighted average. This does not identify institutional accumulation or imply that all shareholders were profitable. More importantly, final VWAP includes trades that had not yet happened in the morning. Using it to evaluate a morning decision would introduce future information. Daily OHLC data cannot reconstruct the complete recovery path, the timing of a VWAP crossing or the sequence of intraday consolidations. ### 6. Observable conditions that would weaken the thesis JPY 13,750 was the high on both 16 and 19 October. A future move above it followed by acceptance would be more informative than a momentary break. Failure to retain the previous close of JPY 13,375 would weaken the observed recovery. Persistent trading below the 19 October low near JPY 12,700 would require reassessing a short-term stabilization thesis. These are observation references, not automatic trading instructions. \[P1\] For the next session, I would use that session’s evolving VWAP and synchronized comparison returns. The 19 October final VWAP remains a historical reference, not tomorrow’s VWAP. A durable business thesis additionally needs disclosed evidence on pricing, volumes, margins and cash flow. ### TokyoBrainie’s interpretation The useful question is not only whether the news was negative, but whether prices continued falling after the opening auction had incorporated that news. Kioxia opened sharply lower and finished above its previous close. That is an observable recovery. It does not establish that the underlying risks have been resolved or that the longer decline has ended. I treat this as a study in separating news, prices and causal explanations. The recovery is established; the dominant buyer and catalyst remain unresolved.

Original Notion archive