インテル入ってる？——入ってないから帰ってくれないか？？｜AI相場、未来の建設費を問い直す朝
Reuters’ final US-session report changes the status of the evidence. The supplied heat maps had already shown severe stock-level losses, but the official close now c
Reuters’ final US-session report changes the status of the evidence. The supplied heat maps had already shown severe stock-level losses, but the official close now confirms that this was not only an Intel event: the Nasdaq Composite fell 2.86%, the S&P 500 1.82%, and the PHLX semiconductor index 7.2%. Intel lost 28.6%, AMD 10.9%, Nvidia 4.2%, while Dell and HP fell 10.8% and 9.6%. Nasdaq decliners outnumbered advancers by roughly three to one.
This strengthens the near-term risk to Tokyo. My previous reading was that Monday could move from indiscriminate semiconductor selling toward company-level selection after strong Tokyo Electron and Kioxia disclosures. That selection can still occur, but it must now absorb a much larger sector beta shock. Tokyo Electron’s positive PTS close-to-close comparison and buyback authorization are evidence for relative resilience, not protection from a weak opening. Kioxia’s PTS decline is a separate sign that a strong operating result can still lose to expectations and positioning.
The rate channel provides an important counterpoint. US Treasuries rallied and the 10-year yield moved below 5%, while the yen was around 157.5 per dollar. This is not a selloff caused by higher discount rates. It is a repricing of AI cash conversion, execution and crowded exposure, with lower yields failing to support technology shares.
For Tokyo’s next session, the key test is not merely whether chip shares gap down. It is whether Tokyo Electron, Advantest, Kioxia, Ibiden and related suppliers recover their opening price and VWAP with breadth and turnover. If strong company guidance and buybacks produce sustained relative strength while SOX stabilizes, selection is working. If even the strongest reporters fail to recover and TOPIX breadth weakens, the market is still trading the global theme rather than company fundamentals.
“Intel inside? If it isn't, could you please leave?” TokyoBrainie’s joke captures the mood. “Divine punishment” is our satirical headline language, not a causal explanation for a stock move. Nor does this session alone establish that an earlier joke predicted the market.
The useful question is what investors are refusing to pay for. Three different problems can meet in the same red rectangle: a company’s delivery problem, an infrastructure owner’s investment-payback problem, and a portfolio’s exposure problem. The common financial table below provides the primary-source evidence without mixing reporting periods. The market table records the supplied screen, rather than claiming independently certified closing prices.
First, evaluate Intel’s demand mix, capacity and profitability separately. An expanding opportunity is not automatically deliverable revenue; deliverable revenue is not automatically attractive margin. Weakness in one end market does not establish the disappearance of every other end market. Equally, management’s confidence is a claim to test, not a substitute for execution. A turnaround valuation needs a credible bridge from demand to shipment to cash.
Second, evaluate Amazon as both a seller of computing services and an owner of expensive infrastructure. The investor underwriting the infrastructure must consider utilization, pricing, power, depreciation, maintenance and replacement. A supplier may collect payment before the customer earns a satisfactory lifetime return. These businesses therefore need not deserve the same valuation multiple at the same moment.
Negative free cash flow is a funding and return question; it is not, by itself, evidence that customers have vanished or that a company is insolvent. Check the period, definition, liquidity and obligations. Do not turn a trailing-twelve-month figure into a single-quarter loss, and do not label all group capital expenditure “GPU spending.”
Third, price contagion can exceed immediate earnings exposure. Selling an industry basket may affect businesses whose customer mix and order visibility differ sharply. The maps show different outcomes across technology and some green defensive pockets. They do not identify the sellers, prove forced liquidation or measure net flows.
Read the four panels carefully. Market-cap sizing shows the weight of listed equity; price-times-volume sizing shows trading activity, not one-way selling. The electronic-technology drilldown includes Apple and is not a pure semiconductor index. Relative volume is activity against a reference level, not a return, a buy signal or an institutional accumulation detector. A stock can fall on below-reference volume. The ratios should not be assigned a precise lookback without checking the platform’s settings.
The cropped five-minute charts add path, not a new daily-return table. A late rebound can coexist with a severe daily decline. Short covering, closing auctions and long-only buying are possible explanations, but the screenshot cannot distinguish them. The small percentages beside candle headers describe individual bars; they must not be reported as the session’s move. The Treasury panel likewise cannot establish a daily yield change from its last candle.
For Tokyo’s next session, separate earnings from discount rates. Our recent Tokyo Electron and Kioxia dispatches remain useful company-level reference points; strong operating news does not guarantee protection against a global repricing. Test equipment, wafer-fab equipment, memory, interconnects and power infrastructure should be examined through different order books. The observation checklist below is a research agenda, not a list of automatic buys or sells.
What would change this reading? Sustained relative strength accompanied by unchanged orders and investment plans would support the view that indiscriminate selling overshot fundamentals. Broad order cuts, delayed projects or deteriorating cash conversion would support a more persistent correction. One green opening candle establishes neither outcome.
The punchline is not that technology stopped mattering. It is that a promising future still needs a financing plan, a delivery schedule and a return on the capital used to build it.
Source method: Intel and Amazon releases were inspected. The supplied CNBC pages and original Reuters article could not be read in full; they are linked as the user’s news leads, not represented as fully verified text. Screenshot observations are attributed to TokyoBrainie’s TradingView captures. Interpretations above are ours.
The shared table records Kabutan’s displayed Japan NEXT PTS trades, captured around 05:34–05:35 JST. Tokyo Electron remained above its TSE close but below its overnight VWAP; Kioxia was below both. This divergence supports examining the companies separately. A positive close-to-PTS comparison does not erase deterioration within the overnight session.
The timestamps differ, so this is not a synchronized comparison after the same final US-market information. PTS prices do not determine the next TSE opening. Earnings assessment and overseas risk aversion may overlap, but their individual contributions cannot be identified here. Reassess against the regular-session close, fresh intraday VWAP and volume when Tokyo reopens. The page’s liquidity label does not guarantee PTS depth; the margin-balance panel is dated October 23 and cannot establish liquidation during this selloff. Sources: the two Kabutan links above; percentages calculated from the table.
