> ## Content Index
> Fetch the complete content index at: https://tokyobrainie.ghost.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# 10月29日 朝刊｜メモリーは稼ぐ、AIは選んで使う——Fed据え置き後の東京市場｜Memory Earns. AI Buyers Choose.
- URL: https://tokyobrainie.ghost.io/news-cutting-room-407bc361-49fcf4f5-en/
- Published: 2026-10-30T03:00:00.000Z
- Updated: 2026-10-30T03:00:00.000Z
- Description: The Nikkei closed at 63,290.11, down 2.12%, after falling as much as 3%. TOPIX lost 0.95% to 3,965.07. Reuters reported that the market reaction to the BOJ decision
- Author: TokyoBrainie
- Tags: News Cutting Room, Signal Room, English, #Import 2026-11-11 08:54

[English](https://tokyobrainie.ghost.io/news-cutting-room-407bc361-49fcf4f5-en/) · [简体中文](https://tokyobrainie.ghost.io/news-cutting-room-407bc361-49fcf4f5-zh-hans/) · [日本語](https://tokyobrainie.ghost.io/news-cutting-room-407bc361-49fcf4f5-ja/)

The Nikkei closed at 63,290.11, down 2.12%, after falling as much as 3%. TOPIX lost 0.95% to 3,965.07\. Reuters reported that the market reaction to the BOJ decision and faster ETF sales was muted; the selloff remained concentrated in AI and semiconductor names. This changes the causal ranking: the first-day decline was mainly a reassessment of AI capital returns and valuation, not evidence of immediate BOJ ETF liquidation pressure.

Post-close company information now creates a two-way test for Monday. Tokyo Electron forecast FY2027 operating profit of ¥1.169 trillion, up 87.1%, and authorized a buyback of up to ¥150 billion / 25 million shares. Kioxia’s official filing confirmed Jul-Sep revenue of ¥2.480 trillion and strong AI data-center demand, but its implied Oct-Dec net-profit forecast of about ¥1.36 trillion remains below the cited QUICK consensus. Strong demand and demanding expectations can coexist.

The counter-test is Monday’s opening and first hour: whether Tokyo Electron’s earnings and buyback reverse Friday’s 5.2% fall, and whether Kioxia can hold despite the forecast gap. A broad rebound without narrowing market breadth would not invalidate the valuation concern.

---

### What changed

The morning test set at 07:33 JST was explicit: a gap-down alone would not prove the end of AI demand, but failure to recover VWAP with turnover would confirm transmission of the US valuation reset. That recovery did not occur. Reuters reported the Nikkei down as much as **3%** and about **2.8%** late in the morning, while the TOPIX lost **1.3%**. Advantest fell **9.1%**, Tokyo Electron **5.6%**, Ibiden **15.2%**, Lasertec **13.8%** and SoftBank Group **6.7%**.

The catalyst was not evidence that AI demand vanished. It was concern that hyperscaler spending and massive cash outflows would not convert into acceptable returns quickly enough. Amazon and Intel's post-close reactions added to the repricing after the US regular session.

### Previous and updated judgment

**Previous judgment:** remain cautious only if the semiconductor group failed to recover VWAP by 09:30; distinguish a valuation shock from a collapse in end demand.

**Updated judgment:** the caution condition was met. Tokyo did not absorb the US shock; index-heavy AI and semiconductor shares amplified it. The structural demand thesis for testing, substrates, memory, networking, power and cooling remains unbroken, but near-term multiples and positioning now require stronger cash-flow proof.

### Tokyo transmission and next checks

The selloff is concentrated enough to punish the Nikkei more than the broader TOPIX, but broad enough to include equipment, substrates and AI-financing proxies. The BOJ's decision adds a second afternoon test: higher-rate expectations and faster ETF disposal could extend pressure on high-duration index heavyweights, while banks may diverge positively.

At the afternoon session and next close, verify whether turnover expands, whether the group recovers the morning low and VWAP, whether the Nikkei/TOPIX gap narrows, and whether US semiconductor futures stabilize. A rebound without improved breadth would be a technical bounce, not full invalidation.

### Sources

- [Reuters｜Nikkei slumps 3% as AI-spending concerns hit chip shares](https://www.reuters.com/world/asia-pacific/japans-nikkei-slumps-3-tech-stocks-tumble-ai-worries-2026-10-30/?ref=tokyobrainie.ghost.io)
- [Bank of Japan｜30 October 2026 policy statement](https://www.boj.or.jp/en/mopo/mpmdeci/mpr%5F2026/k261030a.pdf?ref=tokyobrainie.ghost.io)

---

### What the new data add

US real GDP grew at a **3.4% annual rate** in the third quarter. Real final sales to private domestic purchasers rose **4.9%**. September PCE inflation accelerated to **3.6% year on year**, with core PCE at **3.1%**. Yet real PCE was essentially flat month on month, real disposable income fell **0.3%**, and the saving rate was **3.8%**.

The equity reaction was selective rather than a full risk-off event. The Nasdaq fell **1.37%** and the SOX **1.4%**, while the Russell 2000 rose **1.0%**. Alphabet fell 12.6%, and Amazon later fell about 6.5% after hours despite strong AWS growth.

### Tokyo transmission

The demand evidence supports testing, advanced substrates, memory, optical links, power and cooling. The inflation and cash-flow evidence raises discount rates and narrows the group of companies investors will fund at high multiples.

Today's BOJ decision adds a second axis. A hold with a clear December-tightening signal could strengthen the yen and lift short JGB yields, amplifying pressure on high-PER exporters and semiconductor shares. A cautious hold could leave USD/JPY near the intervention boundary instead.

### Pre-open test

Do not treat a semiconductor gap-down as proof that the AI demand cycle has ended. By 09:30, check whether Advantest, Ibiden, Tokyo Electron, SCREEN, Kokusai Electric, Disco and Kioxia recover VWAP with turnover. Also separate bank strength driven by the yield curve from broad market strength.

### Sources

- [BEA｜Q3 2026 GDP advance estimate](https://www.bea.gov/news/2026/gdp-advance-estimate-3rd-quarter-2026?ref=tokyobrainie.ghost.io)
- [BEA｜September 2026 Personal Income and Outlays](https://www.bea.gov/news/2026/personal-income-and-outlays-september-2026?ref=tokyobrainie.ghost.io)
- [Reuters｜US market close, 29 October](https://www.reuters.com/business/nasdaq-futures-slide-mixed-big-tech-results-test-ai-trade-2026-10-29/?ref=tokyobrainie.ghost.io)
- [BOJ｜2026 monetary-policy meeting schedule](https://www.boj.or.jp/en/mopo/mpmsche%5Fminu/?ref=tokyobrainie.ghost.io)

---

### What changed

Amazon reported after the US close that third-quarter net sales rose 12% to **USD 201.9 billion**. AWS revenue grew **39% to USD 46.0 billion**, its fastest growth in nearly five years, and AWS operating income increased to **USD 17.8 billion** from USD 11.4 billion.

The cash-flow test was much weaker. Amazon's trailing-twelve-month operating cash flow rose 29% to **USD 168.3 billion**, but property and equipment purchases net of proceeds and incentives reached **USD 210.657 billion**. Free cash flow fell to **negative USD 42.357 billion** from positive USD 14.788 billion a year earlier. Amazon maintained approximately **USD 220 billion** of 2026 cash capital expenditure. Fourth-quarter sales guidance of USD 240–247 billion had a midpoint below the USD 244.51 billion LSEG consensus. Reuters reported at **06:21 JST** that Amazon shares were down nearly 7% in extended trading. This was an early post-results reaction, not the next regular-session close.

Intel reported revenue of **USD 16.5 billion, up 21%**. Data Center and AI revenue rose **78% to USD 7.3 billion**, while Intel Foundry revenue rose 58% to USD 6.7 billion; foundry revenue includes intersegment transactions. Client Computing and Physical AI revenue fell 5% to USD 8.1 billion.

Intel's fourth-quarter guidance was the problem: revenue of **USD 15–16 billion** versus the USD 17.07 billion LSEG estimate, and adjusted EPS of USD 0.25 versus USD 0.43\. Intel said the PC slowdown would continue at least through the first half of 2027\. An explosion at its Kiryat Gat fab is expected to reduce revenue by roughly USD 300 million in both the fourth quarter and first quarter. Reuters reported Intel shares down about **18%** after hours and negative spillover across semiconductor shares.

During the regular session, the Nasdaq fell **1.38%** and the SOX fell about **1.4%**. The US 10-year yield fell to around **4.999%** during trading and dropped further after the results, while USD/JPY was approximately **157.87**. Lower yields reflected a risk-off bond bid; they did not prove that AI financing risk had disappeared.

### Previous and updated judgment

**Previous judgment:** AI-server demand was reaching Japanese supplier earnings and capacity plans, while the market would separate demand from cash conversion and execution.

**Updated judgment:** Amazon and Intel confirm that the demand side remains real. AWS growth, Intel data-center growth and Amazon's USD 220 billion capital plan support data-center demand. Yet the shareholder test has become stricter. Amazon's negative USD 42.4 billion trailing free cash flow and Intel's weak guidance produced severe price reactions despite strong AI-related revenue.

The selection mechanism is now visible across four hyperscaler and semiconductor reports: Microsoft converted demand into positive free cash flow; Alphabet and Amazon did not; Intel's data-center growth could not offset PC weakness, supply constraints and disruption risk.

### Tokyo transmission

For Advantest and Ibiden, the structural demand evidence remains intact. For Tokyo Electron, SCREEN, Kokusai Electric, Disco and materials suppliers, the signal is mixed: Amazon's capital plan supports equipment demand, while Intel's weak outlook and foundry execution risk can pressure the sector at the open.

The largest immediate risk is a high-expectation unwind. Tokyo closed strongly on 29 October, but the SOX decline and Intel's post-close collapse arrive before the BOJ decision. If the BOJ signals a December hike and the yen strengthens, semiconductor valuation pressure could compound.

### Counterevidence and next checks

1. Whether Advantest and Ibiden hold their previous close or repeat a high-open reversal.
2. Whether Tokyo Electron, SCREEN, Kokusai Electric and Disco trade with Amazon's capex or Intel's weak guidance.
3. The BOJ's official decision, Outlook Report, vote split and Governor Ueda's press conference.
4. Two- and five-year JGB yields and USD/JPY after the BOJ announcement.
5. Whether Amazon explains the USD 42.4 billion free-cash-flow deficit as a peak investment period with measurable returns.
6. Intel's Kiryat Gat restart, wafer recoverability and the two-quarter USD 600 million revenue effect.
7. Confirmed Ministry of Finance intervention. A verbal warning is not intervention.

The cautious view would be disproved if Japanese suppliers absorb the US weakness, Amazon's stock stabilizes after the call and the BOJ does not lift short yields or the yen. It would be reinforced by a semiconductor gap-down that fails to recover by 09:30.

### Sources

- [Amazon｜Q3 2026 official results](https://ir.aboutamazon.com/news-release/news-release-details/2026/Amazon-com-Announces-Third-Quarter-Results/?ref=tokyobrainie.ghost.io)
- [Reuters｜Amazon cash burn and outlook](https://www.reuters.com/business/retail-consumer/amazon-projects-fourth-quarter-sales-below-estimates-2026-10-29/?ref=tokyobrainie.ghost.io)
- [Intel｜Q3 2026 official results](https://www.intc.com/news-events/press-releases/detail/1786/intel-reports-third-quarter-2026-financial-results?ref=tokyobrainie.ghost.io)
- [Reuters｜Intel outlook and factory disruption](https://live.euronext.com/en/financial-news/intel-signals-slower-growth-pc-market-remains-under-pressure?ref=tokyobrainie.ghost.io)
- [Reuters｜Intel after-hours spillover to semiconductors](https://www.marketscreener.com/news/weak-quarter-for-intel-sends-shivers-through-cooling-chip-stocks-ce785fdadc81f720?ref=tokyobrainie.ghost.io)

---

### What changed

The Tokyo close partly disproved the midday warning. The Nikkei finished at **64,659.25, up 1.10%**, and TOPIX recovered from a 0.30% morning decline to **4,003.28, up 0.29%**. Among Nikkei constituents, 132 rose and 92 fell. Advantest closed **6.01% higher**, Tokyo Electron reversed its morning loss to finish **2.35% higher**, Kioxia gained 1.15%, SCREEN 0.37% and Kokusai Electric 1.25%. Furukawa Electric rose 4.75%. The rebound was broader than at midday, though Disco was nearly flat, Renesas fell 4.63% and the Growth Market 250 lost 0.87%.

The rates market did not price an urgent BOJ hike. Reuters reported that the two-year JGB auction's bid-to-cover ratio improved to **4.03 from 3.89**. The two-year yield fell 0.5 basis point to **1.895%**, the five-year yield was unchanged at 2.265%, and the ten-year yield added 0.5 basis point to 2.83%. This was market evidence of demand for short government bonds, not a BOJ policy decision.

At **15:20 JST**, Ibiden reported first-half sales of **JPY 262.546 billion, up 34.3%**, and operating profit of **JPY 60.116 billion, up 84.6%**. Electronics operating profit rose 96.7%. Ibiden raised its full-year sales forecast from JPY 550 billion to **JPY 565 billion** and operating profit from JPY 127 billion to **JPY 133 billion**. It said higher-than-expected orders for high-value-added AI-server products and a weaker yen supported the revision, and it expects robust AI-server, general-server and switch-IC demand.

Ibiden separately approved approximately **JPY 151 billion** of investment at its Gama plant and another overseas plant for high-performance CPU package substrates. Mass production is planned to begin sequentially in fiscal 2029\. The company explicitly said the current-year earnings impact would be immaterial. This is capacity commitment, not near-term profit.

### Previous and updated judgment

**Previous judgment at midday:** record memory profit was not broadening the rally; the positive Nikkei depended on Advantest while TOPIX and breadth weakened.

**Updated judgment:** that warning was partly disproved by the close. Advantest retained its gain, Tokyo Electron turned positive, TOPIX recovered and Nikkei breadth improved. Ibiden then supplied company-specific evidence that AI-server package demand is reaching Japanese earnings and long-dated capacity plans.

The all-clear is still not warranted. Growth stocks remained weak, not every chip name participated, and Ibiden's new capacity starts in fiscal 2029\. A strong two-year auction also shows that bonds did not confirm rapid BOJ tightening, but Friday's actual decision can supersede that signal.

### Tokyo transmission

Testing and advanced packaging now have both earnings and capacity evidence: Advantest for testing, Ibiden for high-performance substrates. Furukawa Electric's 4.75% rise also shows that the afternoon recovery spread into parts of optical and electrical infrastructure.

For the next session, the immediate positive is supplier-specific. The major macro risk is the BOJ: an unchanged policy rate with a December-tightening signal could still lift shorter yields and strengthen the yen. Amazon and Intel results before the Tokyo open will test whether AI demand and cash conversion remain compatible.

### Counterevidence and next checks

1. The BOJ decision, Outlook Report, vote split and Governor Ueda's press conference on 30 October.
2. Two- and five-year JGB yields after the decision, not before it.
3. USD/JPY around 157.8 and any confirmed Ministry of Finance intervention.
4. Whether Ibiden's orders are backed by firm volume commitments, price terms and customer diversification.
5. The execution burden of JPY 151 billion of new investment and the fiscal-2029 start date.
6. Amazon's AWS growth, capital expenditure and free cash flow; Intel's data-center and foundry outlook.
7. Whether Advantest, Ibiden and Tokyo Electron hold gains after the BOJ rather than only after earnings.

The constructive revision would fail if the BOJ triggers a sharp rise in short yields and yen appreciation, if US hyperscaler results renew cash-burn concerns, or if Ibiden's investment raises capital intensity without matching contracted demand.

### Sources

- [JPX｜29 October official index values](https://www.jpx.co.jp/english/markets/indices/realvalues/?ref=tokyobrainie.ghost.io)
- [Reuters｜Strong two-year JGB auction ahead of the BOJ](https://www.reuters.com/world/asia-pacific/japan-bonds-steady-ahead-boj-decision-strong-auction-lends-support-2026-10-29/?ref=tokyobrainie.ghost.io)
- [Ibiden｜Q2 FY2026 results and disclosures](https://www.ibiden.com/ir/?ref=tokyobrainie.ghost.io)
- [Ibiden｜JPY 151 billion package-substrate investment](https://www.ibiden.com/ir/items/koteishisansyutoku2026Q2.pdf?ref=tokyobrainie.ghost.io)
- [Ibiden｜Forecast revision and dividend](https://www.ibiden.com/ir/items/gyousekishuusei2026Q2.pdf?ref=tokyobrainie.ghost.io)

---

### What changed

At **10:00 JST on 29 October**, Samsung Electronics released its third-quarter results. Consolidated revenue was a record **KRW 195.3 trillion**, up 14% quarter on quarter, and operating profit reached **KRW 107.4 trillion**. The Device Solutions division generated **KRW 147.1 trillion** of revenue and **KRW 106.5 trillion** of operating profit. Samsung said memory revenue and profit set records, HBM4 sales expanded significantly, and 2027 AI-infrastructure investment could reach a historic level while demand growth exceeds industry supply. These are company statements, not guaranteed industry outcomes.

The price response was less constructive. Reuters reported at **11:28 JST** that Samsung shares were down more than 4% and the KOSPI was down 2.6%, even after Samsung's record result. Alphabet's post-close decline widened to about 10% and Nasdaq futures fell about 1%. In Tokyo at the **11:30 JST** break, the Nikkei was **64,131.28, up 177.09 points or 0.28%**, but TOPIX was **3,979.89, down 0.30%**. Prime Market decliners numbered **1,022** against **478** advancers.

Advantest rose **5.41%** to JPY 36,820 and added approximately **407.76 points** to the Nikkei. Its contribution was 2.3 times the index's total gain; excluding Advantest, the Nikkei would have been about 231 points lower. Kioxia rose 3.56% and Ibiden 2.68%, while Tokyo Electron fell 0.88%. USD/JPY was around **157.68** at 11:41 JST. No intervention was confirmed.

### Previous and updated judgment

**Previous judgment:** AI and memory demand are real, and Tokyo suppliers can benefit, but high yields and cash conversion will split winners from financing-dependent buyers.

**Updated judgment:** the demand thesis is now supported by Samsung's official results as well as SK hynix and Advantest. Yet the market has raised the hurdle. Record profit did not prevent Samsung from falling, and Tokyo's positive headline index depended on one stock. The key risk has shifted from “is demand real?” to “how much perfection is already priced, and can earnings breadth catch up?”

This is not evidence that semiconductor demand has collapsed. It is evidence that stronger fundamentals are no longer sufficient for a broad re-rating when expectations, discount rates and concentration are high.

### Tokyo transmission

Advantest's result was accepted, so testing remains the clearest positive pocket. Kioxia and Ibiden also retained gains. The weak TOPIX and two-to-one ratio of decliners show that the positive read-through did not extend across the market.

The split matters for the next session. If Amazon and Intel confirm demand but their shares also fall, Tokyo may continue to reward only suppliers with company-specific earnings upgrades. If the BOJ keeps rates unchanged but signals a December move, the yen and shorter JGB yields could add another valuation headwind.

### Counterevidence and next checks

1. Whether Advantest closes above its morning VWAP and whether Tokyo Electron, Disco and SCREEN participate.
2. Whether TOPIX and advance/decline breadth recover in the afternoon.
3. Whether Samsung stabilizes despite record profit and whether SK hynix retains its gain.
4. Amazon's AWS growth, capital expenditure and free cash flow, plus Intel's data-center and foundry outlook, due before Tokyo's 30 October open.
5. The BOJ decision, Outlook Report language, vote split and Governor Ueda's comments on 30 October.
6. The two-year JGB auction, 2- and 5-year yields, and USD/JPY near 158.
7. Any confirmed Ministry of Finance intervention. Verbal warnings alone are not intervention.

The caution would be disproved by a broader close in Tokyo, stabilization in Samsung and Alphabet, and supplier gains spreading beyond a few index-heavy names. It would be reinforced if strong earnings keep producing negative price reactions and breadth deteriorates further.

### Sources

- [Samsung Electronics｜Q3 2026 official results](https://news.samsung.com/global/samsung-electronics-announces-third-quarter-2026-results?ref=tokyobrainie.ghost.io)
- [Samsung Investor Relations｜29 October earnings call](https://www.samsung.com/global/ir/?ref=tokyobrainie.ghost.io)
- [Reuters｜Asian markets, Samsung, Alphabet and Nasdaq futures](https://www.reuters.com/world/china/global-markets-global-markets-2026-10-29/?ref=tokyobrainie.ghost.io)
- [Tokyo morning market data｜11:30 JST breadth and contribution](https://atpages.jp/tokyo-am-nikkei-177-advantest-408-banzuke-2026-10-29/?ref=tokyobrainie.ghost.io)

---

Reuters reports DOJ inquiries into whether NVIDIA’s proposed Hugging Face acquisition could harm competition; CNBC also reports the scrutiny. This is not a finding of wrongdoing or an announced decision to block the deal.

NVIDIA’s September announcement puts the deal at USD 12.9303 billion. Its SEC filing separates approximately USD 11.9 billion for stockholders from up to USD 1 billion in employee retention equity, with closing expected in the first half of 2027 subject to approvals. NVIDIA promises continued support for other chips and no requirement to use NVIDIA compute.

**My interpretation:** the question is whether control of an important model-distribution platform could disadvantage rival hardware or services through access, integration or commercial terms. These are risks to examine, not conduct established by this report. Open model availability alone would not settle whether commercial access remains neutral.

**Tokyo:** add regulatory execution to the existing demand, cash-flow and financing tests. This report does not establish Japanese equipment-order cancellations or a decline in semiconductor demand. Watch disclosed approval conditions, closing timing and practical support for competing accelerators before changing supplier earnings assumptions.

---

### What changed

At **03:00 JST on 29 October**, the Federal Reserve kept the federal-funds target range at **3.75%–4.00%** in a **10–2 vote**. Lisa Cook and Beth Hammack preferred a 25-basis-point increase. The official statement upgraded economic activity to a strong pace and said inflation remained elevated. This was a pause, not a declaration that tightening was over.

The market did not treat the hold as a broad easing signal. The US 10-year Treasury yield ended near **5.045%**, about 5 basis points higher, while the two-year yield fell to roughly **4.72%**. The curve steepened. USD/JPY was approximately **157.87** in Reuters' post-Fed currency update. No Japanese intervention was confirmed.

Microsoft and Alphabet then reported after the US close. Their cloud businesses both grew strongly, but cash conversion diverged sharply.

Microsoft reported quarterly revenue of **USD 92.3 billion**, up 19%, and Azure revenue of **USD 33.17 billion**, up 48%. Free cash flow reached **USD 18.6 billion**, well above the Visible Alpha estimate of USD 8.54 billion. Capital expenditure including finance leases was **USD 45.2 billion**, below the company's roughly USD 50 billion indication, partly because third-party capacity deliveries were delayed.

Alphabet reported revenue of **USD 127.6 billion** and Google Cloud revenue of **USD 30.71 billion**, up 103%. Yet quarterly capital expenditure reached **USD 51.36 billion** and free cash flow was **negative USD 9.07 billion**, compared with negative USD 5.86 billion in the previous quarter. Alphabet shares fell about **8%** in extended trading. The company also disclosed equity issuance and USD 28.7 billion of senior unsecured notes during the quarter.

### Previous and updated judgment

**Previous judgment:** AI and memory demand were real, but the price of capital and each buyer's ability to turn investment into cash would determine shareholder outcomes.

**Updated judgment:** the split is now visible in reported cash flow. Microsoft demonstrated that strong cloud growth can coexist with positive free cash flow. Alphabet demonstrated that cloud revenue can more than double while capital expenditure overwhelms quarterly operating cash generation. This is not an argument that Alphabet's technology or cloud demand is weak. It is evidence that revenue growth and shareholder cash return are separate tests.

The Fed's two hawkish dissents and a 10-year yield above 5% reinforce that distinction. The cost of financing did not fall simply because the policy rate was left unchanged.

### Tokyo transmission

For Advantest, Ibiden and other testing or advanced-packaging suppliers, strong cloud and AI infrastructure spending supports the demand narrative. Advantest's own raised guidance and buyback remain the strongest company-specific evidence.

For Tokyo Electron, Disco, SCREEN, Kioxia, Fujikura, Furukawa Electric and Sumitomo Electric, the read-through is conditional on customer mix, actual orders, delivery capacity and margins. Microsoft delayed some third-party capacity deliveries; that is a reminder that capex guidance does not translate into identical shipment timing for every supplier.

The risk to the Tokyo open is a split tape: semiconductor suppliers can benefit from demand, while expensive growth stocks face a higher discount rate and hyperscaler cash-burn concerns. A high Nikkei opening without broader TOPIX participation would repeat the concentration problem seen on 28 October.

### Counterevidence and next checks

1. Whether Advantest holds its post-results gain at **09:30 and the close**, not only at the opening auction.
2. Whether Tokyo semiconductor strength broadens beyond a few high-weight names.
3. Whether the US 10-year yield remains above 5% during Asian trading.
4. Whether Alphabet explains the cash burn as temporary capacity timing or a persistent funding need.
5. Microsoft's second-quarter Azure outlook and whether delayed third-party deliveries shift rather than reduce spending.
6. USD/JPY near 158, Ministry of Finance language and any confirmed intervention.
7. The BOJ meeting on 29–30 October and Amazon's results on Japan's 30 October morning.

The constructive thesis would weaken if the supplier rally fades despite strong demand, long yields rise further, or hyperscaler cash generation continues to lag investment. It would strengthen if orders broaden across the Japanese supply chain while free cash flow remains positive and financing spreads stabilize.

### Sources

- [Federal Reserve｜28 October FOMC statement](https://www.federalreserve.gov/newsevents/pressreleases/monetary20261028a.htm)
- [Reuters｜FOMC decision and market reaction](https://kelofm.com/2026/10/28/fomc-holds-rates-steady-markets-mostly-follow-suit/?ref=tokyobrainie.ghost.io)
- [Reuters｜Dollar and USD/JPY after the Fed](https://live.euronext.com/en/financial-news/dollar-falls-after-fed-holds-rates-steady?ref=tokyobrainie.ghost.io)
- [Microsoft｜FY2027 Q1 official results](https://www.microsoft.com/en-us/Investor/earnings/FY-2027-Q1/press-release-webcast?ref=tokyobrainie.ghost.io)
- [Reuters｜Microsoft cloud revenue and free cash flow](https://www.schwab.wallst.com/schwab/Prospect/research/stocks/news.asp?doc=3200-nL3N46E2B2-7&symbol=MSFT&ref=tokyobrainie.ghost.io)
- [Alphabet｜Q3 2026 results](https://www.publicnow.com/view/ED018F4AAE12FB3B1BD1E9649EA9BF381E69B00F?ref=tokyobrainie.ghost.io)
- [Reuters｜Alphabet cash burn and extended-trading reaction](https://www.schwab.wallst.com/schwab/Prospect/research/stocks/news.asp?doc=3200-nL3N46E2AX-6&symbol=MSFT&ref=tokyobrainie.ghost.io)

---

## The morning view

For Tokyo on 29 October, the strongest connection is between memory profitability, cloud monetization, the price of capital and control over technology. These are separate tests. Strong AI demand does not guarantee every supplier’s margin, every operator’s cash return or every shareholder’s outcome.

The common data table below records the key readings. It separates official company results, a policy decision, proposals and reported litigation. Numbers are not live quotations.

## 1 | A Fed pause, not an all-clear

The Fed has already announced its decision; yesterday’s previews are no longer the event. The unchanged range and the two dissents favoring a hike matter together.\[1\] My interpretation is that the pause does not remove inflation or discount-rate risk. Policy rates, Treasury yields, corporate spreads and USD/JPY can move differently.

For Japanese banks, examine funding costs, bond valuations and credit costs alongside lending margins. For exporters and expensive growth stocks, watch the yen and longer yields rather than treating “hold” as a universal buy signal.

## 2 | Real demand: memory and cloud

SK hynix’s release supports a memory-demand thesis, while its HBM4 plans and long-term agreements provide company-stated visibility—not guaranteed future earnings.\[2\] Its Kioxia investment gains and related shareholder distributions should not be confused with recurring memory profit.

Microsoft and Alphabet have also reported; the forecast stage is over.\[3\]\[4\] Their cloud growth is evidence worth studying, but the categories are not identical: Alphabet identifies TPU systems sales as a contributor. Do not compare growth rates as though the product mix were the same. Alphabet’s cloud expansion and lower consolidated EPS also coexist; revenue alone does not explain shareholder earnings.

For Tokyo, distinguish HBM/DRAM, NAND and testing. Kioxia is a NAND research exposure, not a direct substitute for an HBM producer. Advantest, Tokyo Electron and SCREEN have different customer mixes, capacity exposures and order timing. Demand at one manufacturer cannot be copied into every supplier’s earnings estimate.

## 3 | SAP: cheaper work can mean more AI

Reuters corrected its SAP report: cheaper-model substitution contributes to a broader efficiency-gain target of more than €2 billion; it does not alone generate that total.\[5\] It also reports that SAP expects overall AI expenditure to rise with adoption.

The investment question is cost per successfully completed task, not which model has the most prestigious name. Lower unit costs can help adoption and application margins while pressuring some model providers’ pricing. Whether total compute demand rises depends on how much usage expands.

The funding article belongs beside this, not above it as a verdict on the whole technology.\[6\] Equity issuance can dilute existing holders; debt brings refinancing and covenant risks. A customer and supplier financing each other does not by itself prove fraud, but it warrants checking independent cash demand and payment obligations.

## 4 | China: the draft and the live registration requirement differ

MOFCOM’s announcement covers proposed AI, biomedicine and photovoltaic technology additions, while imposing interim registration on covered exports.\[7\] Already-public technology is treated separately. An opinion period is not the same as a fully implemented final catalogue.

For Japanese projects, separate access to a service from technology transfer, licensing and foreign ownership. Review the actual technology and transaction; “China AI” is too broad a legal category. Do not turn a technology-export measure into a claim that Japanese equipment sales to China have all been banned.

## 5 | Margin checks beyond AI

Unitree’s reported profit decline is a check on robotics economics, not proof that robotics has no future.\[8\] For FANUC, Yaskawa and component suppliers, ask about production volumes, customer orders and margins rather than humanoid headlines.

Logitech illustrates why a profit beat can coexist with concern about future margins.\[9\] Rising component costs may benefit some suppliers while squeezing their customers. Unilever’s official volume/price split is a useful consumer comparison; it does not establish that every Japanese household-products company will match it.\[10\]

## 6 | Other countries and risks worth keeping

Britain has announced a budget, replacing the preview. The bank surcharge change in the table is a surcharge, not the total corporation-tax rate.\[11\] Watch gilt issuance, yields, sterling and the implications for exposed businesses. A UK tax change does not mechanically apply to Japanese banks.

Jio’s listing is relevant to India’s digital-capital market, but a debut premium is not a forecast of sustainable returns.\[12\] It is a secondary research item rather than the main Tokyo opening catalyst.

The Meta/Snap case concerns allegations and proceedings, not an established finding of liability in this trial.\[13\] Watch product requirements, legal costs and business-model consequences; do not assign a damages bill to the whole AI industry.

The 23 October ECB preview is background for the 29 October meeting, not its outcome. The BOJ meets on 29–30 October; Amazon’s US 29 October results correspond to Japan’s 30 October morning.\[14–16\]

## Tokyo checklist and countercase

Observe semiconductor leadership after the opening, breadth relative to the indices, same-time volume, USD/JPY and yields. For optical and power infrastructure names such as Fujikura, Furukawa Electric and Sumitomo Electric, require actual orders, delivery capacity and margin evidence.

Constructive: earnings support the demand thesis, gains broaden and financing conditions remain manageable. Fragile: a high opening fades while yields or credit spreads rise. Countercase: much of the good news was already priced on 28 October, and strong revenue fails to translate into cash.

This is a framework for testing the morning, not a buy list or a forecast of the closing index. Apple’s product preview, Alexa devices and Wendy’s deal report receive lower priority here because their immediate connection to this Tokyo framework is weaker. SAP and China links were each supplied twice and counted once.

**Demand can be strong, capital can be expensive, and margins can diverge—all on the same morning.**

---

[Original Notion archive](https://app.notion.com/p/407bc361ccd281e68845e80749fcf4f5?pvs=204&ref=tokyobrainie.ghost.io)