> ## 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. # Deep Cut 006｜季節は債務を返済しない——CoreWeave「秋の押し」と日経9万円予想の解剖｜Seasons Do Not Service Debt｜季节不会偿还债务 - URL: https://tokyobrainie.ghost.io/news-cutting-room-3fdbc361-1a293519-en/ - Published: 2026-11-10T15:00:00.000Z - Updated: 2026-11-11T00:15:11.000Z - Description: ［初出日未確認・表示日は移行日］ Seasons Do Not Service Debt On 16 October 2026, Shikiho Online published a premium column by Daiwa Securities strategist Eiji Kinouchi arguing that CoreWeave's credi - Author: TokyoBrainie - Tags: News Cutting Room, Signal Room, Semiconductor Research, Tokyo Dispatch, English, #Import 2026-11-11 08:54, #Undated archive [English](https://tokyobrainie.ghost.io/news-cutting-room-3fdbc361-1a293519-en/) · [简体中文](https://tokyobrainie.ghost.io/news-cutting-room-3fdbc361-1a293519-zh-hans/) · [日本語](https://tokyobrainie.ghost.io/news-cutting-room-3fdbc361-1a293519-ja/) # Seasons Do Not Service Debt On 16 October 2026, Shikiho Online published a premium column by Daiwa Securities strategist Eiji Kinouchi arguing that CoreWeave's credit anxiety was “seasonal.” The visible subtitle attributed the autumn decline in high-yield debt to loss-taking and argued that it was not a problem for AI as a whole.[\[1\]](https://shikiho.toyokeizai.net/news/0/961443?ref=tokyobrainie.ghost.io) The useful half of that claim is obvious: one leveraged neocloud is not the entire AI economy. The unsupported half is the leap from “not systemic” to “seasonal.” Seasonal tax-loss selling and fund-year positioning can amplify a move. They do not establish why a particular issuer was selected for sale, whether its refinancing economics remain viable, or whether the market is merely wrong. ## The Numbers Autumn Cannot Explain Reuters reported roughly USD 50 billion of CoreWeave debt, CDS near 1,000 basis points and an implied five-year default probability near 63%. A follow-up placed CDS around 1,200 basis points and yields on some bonds near 15–16%.[\[3\]](https://www.reuters.com/legal/transactional/coreweaves-ai-contracts-lenders-financing-its-build-out-2026-10-15/?ref=tokyobrainie.ghost.io)[\[4\]](https://www.reuters.com/legal/transactional/private-credit-roundup-coreweaves-debt-stress-tests-private-credits-ai-financing-2026-10-16/?ref=tokyobrainie.ghost.io) CoreWeave's own first-half filing shows the funding equation. Second-quarter revenue was USD 2.575 billion, operating loss USD 49 million, net loss USD 626 million and net interest expense USD 640 million. First-half operating cash flow was positive USD 3.663 billion, but property-and-equipment purchases consumed USD 14.117 billion while financing activities supplied USD 13.985 billion. Net interest expense rose 121% year on year to USD 1.176 billion. Its top three customers accounted for 72% of quarterly revenue.[\[5\]](https://www.sec.gov/Archives/edgar/data/1769628/000176962826000366/crwv-20260630.htm?ref=tokyobrainie.ghost.io) This is a real company with real demand and real infrastructure. That is precisely why the risk is important. Real demand can still be financed badly. A backlog is valuable, but debt service occurs on a calendar while customer deployment, power delivery, construction and utilization can slip. ## The Strong Countercase A fair critique must include the bull case. Nvidia has a commitment covering unsold capacity, major customers have signed long-duration contracts, and CoreWeave reports a very large backlog. Evercore maintained an outperform rating, and other market participants warned against treating CoreWeave as the poster child for all AI.[\[6\]](https://www.reuters.com/business/coreweave-shares-bonds-headed-steep-weekly-losses-ai-borrowing-concerns-mount-2026-10-16/?ref=tokyobrainie.ghost.io) Those facts may support a bond rebound. They do not prove that a 15–16% required yield is a seasonal illusion. The relevant test is whether cash generation begins to fund more of the build-out before external capital becomes too expensive. ## The Nikkei 88,000–94,000 Illustration At roughly 65,324 on 16 October, 88,000 requires about 34.7% appreciation and 94,000 about 43.9%.[\[7\]](https://tradingeconomics.com/japan/stock-market?ref=tokyobrainie.ghost.io) A chart that repeatedly labels advances “30% higher” can illustrate a conditional pattern. It cannot supply a probability. At an unchanged valuation multiple, index earnings must rise by the same 35–44%. If earnings rise 15%, reaching 88,000 still requires multiple expansion of about 17%; if earnings rise 20%, reaching 94,000 requires roughly 20% multiple expansion. Reuters' February strategist poll had a median of 58,500 for end-2026 and 60,750 for mid-2027\. Kinouchi's range is therefore not consensus but an aggressive upside tail.[\[8\]](https://www.reuters.com/world/asia-pacific/nikkei-set-breach-60000-milestone-by-mid-2027-earnings-growth-foreign-inflows-2026-02-24/?ref=tokyobrainie.ghost.io) A tail scenario may become reality. A paid publication should still show the earnings path, valuation assumption, rates, currency, probability and invalidation conditions. ## Was This Nepotism? I found no evidence of nepotism or an improper editorial relationship. Kinouchi is an established technical analyst with a long record, multiple number-one rankings and two QUICK forecast awards.[\[2\]](https://president.jp/growth/authors/pre-eiji-kinouchi.html?ref=tokyobrainie.ghost.io) That makes the broader problem more interesting, not less. Prestige is a reason to read a hypothesis. It is not evidence that the hypothesis is true. ## A Declaration This is not a personal attack and does not allege fraud, misconduct or bad faith. It is a declaration against a method: converting titles into evidence, charts into probabilities, and confidence into a substitute for valuation. Shikiho literally evokes the four seasons. Seasonality suits the masthead. **But seasons do not service debt.** If someone forecasts Nikkei 94,000, ask: How much earnings growth? What multiple? What discount rate? What invalidates the call? Without those answers, the reader is not buying analysis. **The reader is buying a priced belief.** --- [Original Notion archive](https://app.notion.com/p/3fdbc361ccd281ae8d38ff831a293519?pvs=204&ref=tokyobrainie.ghost.io)