跳到主要內容

AI泡沫評測:英特爾暴跌28%,那指單日蒸發1兆美元 | AI Bubble Review: Nasdaq's $1 Trillion One-Day Wipeout

By Kit 小克 | AI Tool Observer | 2026-10-31

🇹🇼 AI泡沫評測:英特爾暴跌28%,那指單日蒸發1兆美元

AI泡沫這個詞,這個週末又被罵上熱搜。2026年10月30日,那斯達克指數單日重挫2.9%,科技股市值一天蒸發超過1兆美元,創下自2025年3月以來最慘的一個月。但在你跟著恐慌之前,先看數字,不要只看標題。

導火線:英特爾、亞馬遜財報踩雷

這波AI泡沫擔憂不是憑空出現,是被兩份財報點燃的:

  • 英特爾股價單日暴跌28.6%,市場對其AI轉型進度失去信心
  • 亞馬遜財報顯示AI資本支出持續攀升,但營收與利潤都不如分析師預期,股價應聲下跌
  • 那斯達克100指數成分股單日蒸發逾1兆美元市值,創下2025年6月5日以來最差單日表現

真正的問題:7250億美元資本支出,回報在哪裡?

比起單日股價波動,更值得關注的是底層結構:Google、微軟、Meta、亞馬遜四大巨頭,2026年合計資本支出預估達7250億美元,比去年暴增77%。投資人開始要求證據——這些錢燒下去,到底換回多少實際營收成長?Meta因為不是雲端服務商,被認為風險最高;微軟Azure成長約40%、Google Cloud成長82%,算是少數拿得出成績單的。

這是泡沫破裂,還是正常修正?

老實說,兩者都有可能,現在下結論都太早。股價修正通常反映「預期落空」,不代表底層技術停滯——Claude、GPT、Gemini這幾個月的模型能力並沒有變差,甚至還在進步。但如果巨頭明年真的縮減資本支出,對開發者最直接的影響會是:API降價速度變慢、新模型發布節奏可能放緩、免費額度可能收緊。

對AI工具使用者的誠實建議

  • 不要因為股價新聞就停用正在幫你省時間的工具,技術價值跟股價是兩件事
  • 但也別把身家壓在單一供應商——多準備一個備案(例如同時熟悉Claude跟開源模型),避免哪天某家公司真的縮編AI部門時措手不及
  • 關注資本支出與財報說明會,比盯著股價本身更有參考價值

AI泡沫會不會破,沒人能保證。但工具好不好用,從來不是股價決定的。好不好用,試了才知道。


🇺🇸 AI Bubble Review: Nasdaq's $1 Trillion One-Day Wipeout

The AI bubble debate flared up again this weekend. On October 30, 2026, the Nasdaq Composite dropped 2.9% in a single day, wiping out more than $1 trillion in tech stock value and capping off the index's worst month since March 2025. Before you panic along with the headlines, look at the numbers first.

What Triggered It: Intel and Amazon Earnings Miss

This round of AI bubble anxiety wasn't abstract — it was lit by two earnings reports:

  • Intel shares plunged 28.6% in a single day as the market lost confidence in its AI pivot
  • Amazon reported rising AI capex but revenue and profit that missed analyst expectations, sending shares lower
  • Nasdaq 100 companies lost over $1 trillion in market cap in one session — the worst single day since June 5

The Real Question: $725 Billion in Capex, Where's the Return?

More important than one bad trading day is the underlying structure: Google, Microsoft, Meta, and Amazon combined are projected to spend $725 billion on AI capex in 2026 — up 77% from last year. Investors are now demanding proof that this spending converts into real revenue. Meta is seen as the riskiest bet since it isn't a cloud provider, while Microsoft Azure (~40% growth) and Google Cloud (82% growth) are the closest thing to a scoreboard so far.

Bubble Pop or Normal Correction?

Honestly, it's too early to call either way. A stock correction usually reflects missed expectations, not a stall in the underlying technology — Claude, GPT, and Gemini haven't gotten worse these past few months; if anything they've kept improving. But if hyperscalers actually cut capex next year, developers would feel it directly: slower API price drops, slower model release cadence, and tighter free tiers.

A Straight Answer for AI Tool Users

  • Don't drop a tool that's genuinely saving you time just because of stock headlines — technical value and stock price are different things
  • But don't bet everything on one vendor either — keep a backup plan (e.g. know both Claude and an open-source model) in case a provider scales back its AI division
  • Watching capex guidance and earnings calls tells you more than watching the stock ticker

Nobody can promise whether the AI bubble will actually burst. But whether a tool is genuinely useful has never been decided by its parent company's stock price. You won't know until you try it.

Sources / 資料來源

延伸閱讀 / Related Articles


AI 工具觀察站 — 每日精選 AI Agent 與工具趨勢
AI Tool Observer — Daily curated AI Agent & tool trends

留言

這個網誌中的熱門文章

Google Ironwood TPU v7 推理專用晶片解析:效能追平 NVIDIA、成本低 44%,AI 晶片戰爭正式開打 | Google Ironwood TPU v7 Explained: Matching NVIDIA Performance at 44% Lower Cost — The AI Chip War Heats Up

Claude Code 實測:AI 幫你寫程式到底行不行? | Claude Code Review: Can AI Really Code for You?

Cursor vs GitHub Copilot vs Claude Code:AI 程式助手大比拼 | AI Coding Assistants Compared: Cursor vs GitHub Copilot vs Claude Code