ASSETCASTING
← All topic comparisons

ASSETCASTING · EDITORIAL COMPARISON

Semiconductor selloffs: separating supply news, HBM demand and selling flows

Supply competition, long-term AI infrastructure plans and short-term prices answer different questions. This comparison keeps their time horizons and evidence separate.

Source review date

Read the connections, not just the headlines

Supply reporting should begin with what a competitor can actually produce. Plans, mass production, customer qualification and shipments are different stages. A technology name or capacity plan alone does not establish a rival’s lost sales or margin compression. Check the original report, product category, production stage and verification date.

Microsoft’s October 12, 2026 Maia deployment announcement is evidence of its published AI infrastructure plans. It can inform a discussion of potential HBM demand, but it does not establish a particular memory supplier’s booked orders, selling prices or profit. It should not be presented as Microsoft’s response to separate reporting about a competitor.

Prices can fall in the short term despite constructive long-term business expectations. Deleveraging, risk reduction and position unwinds are separate mechanisms worth testing. The BIS analysis of August 2024 is historical evidence, not proof that the same trades occurred in 2026. Attributing the current move requires matching-date price, volume and position evidence.

Questions to check across the original videos

The topics below follow the linked learning-note summaries. The comparison and verification questions are ASSETCASTING’s own editorial synthesis, not quotations from the video creators. Read each note alongside its original video.

  • 코스피가 무너진 근본적인 이유 (CXMT 아님) →

    내일은 투자왕 - 김단테 · The video distinguishes CXMT news as a catalyst from underlying explanations, discussing AI investment durability, flows, exchange rates and shareholder returns.

    Verification question: How do concerns about AI investment durability connect in time to the day’s prices, turnover and foreign selling? Separate the speaker’s explanation of causes from directly observed figures.

  • [10월 13일 마감시황] 코스피 매도 사이드카 발동에 반도체 투톱 급락…CXMT 악재만으로 설명 안 되는 하락ㅣ홍선애, 김장열, 박명석 [클로징벨 라이브] →

    삼프로TV 3PROTV · This closing-market discussion considers the CXMT technology announcement, a program-selling sidecar activation, foreign selling and share-buyback flows.

    Verification question: What does the original technology announcement establish, and what was observed in that day’s market? Compare the additional evidence required for flow explanations rather than merging several speakers’ interpretations into one established cause.

  • 중국 CXMT에 시장 흔들릴 때, MS는 차세대 칩에 HBM 때려넣는다 →

    머니네버슬립 · The video connects Microsoft’s custom AI chips with changing HBM demand sources, distinguishing commodity-memory competition from HBM demand.

    Verification question: Do the published deployment plans and memory configuration establish a particular supplier’s orders or profit? Keep undisclosed suppliers separate from the video’s outlook.

  • [홍장원의 불앤베어] 반도체 주가에는 새 도파민이 필요하다 →

    매경 라이브 · The video brings together memory-supply forecasts, several AI announcements and expectations of new demand to discuss market narratives.

    Verification question: What dates and product categories underlie the supply forecasts, and what supports expectations of new demand? Check third-party views against their original sources and separate a need for a new narrative from evidence of actual orders or price recovery.

  • "또 시작이네..." →

    월텍남 - 월스트리트 테크남 · The video discusses 4F² cell structure, hybrid bonding and yield while distinguishing a product-release announcement from mass-production readiness.

    Verification question: What has been disclosed about capacity, speed and release timing versus actual production, qualification and shipments? Do not equate illustrative yield calculations with the company’s independently verified yield.

Claims and evidence

Examples of inferences to test. These are editorial checks, not quotations attributed to the linked videos.

  1. Supply news establishes a specific company’s profit decline.

    What the evidence establishes
    Distinguish the reported stage: plan, production, qualification or shipment.
    What needs a separate check
    Substitution, actual customer switching and revenue or margin effects need separate evidence.

  2. An AI infrastructure announcement guarantees a rebound today.

    What the evidence establishes
    A company announcement establishes a long-term plan, not a short-term price forecast.
    What needs a separate check
    Deployment timing, memory configuration, supply contracts and price recognition are separate checks.

  3. A past carry-trade unwind caused this selloff.

    What the evidence establishes
    The BIS source describes possible transmission mechanisms in the 2024 episode.
    What needs a separate check
    Without direct evidence of matching 2026 positions and flows, it cannot establish the current cause.

What this comparison cannot establish

This is an editorial comparison of evidence layers. It does not calculate contributions to a particular 2026 selloff or independently verify CXMT production or qualification. Strong wording in a video is not converted into an official company position or a demonstrated market cause.

Primary sources and their scope