Everyone Says AI Is a Dot-Com BUBBLE — Urgent Update
所有人都说AI是互联网泡沫——紧急更新
Mag7 at 27× forward PE vs 52–63× during dot-com peak. Burry shorting since 2021 — wrong every year. AI revenue is real: $13B Microsoft, $47.5B Nvidia, $20B Meta. Not a bubble. But concentration risk is real.
Mag7当前27倍前瞻市盈率，互联网泡沫顶峰52–63倍。Burry从2021年做空——每年都错。AI营收真实：微软$130亿、英伟达$475亿、Meta$200亿。不是泡沫。但集中度风险是真的。
The Bear Case — What's Worrying People
看空理由——让人担忧的
⚠️ Concentration Risk
⚠️ 集中度风险
Mag7 = 30–37% of S&P 500 market cap. At the dot-com peak, top 7 = only 19%. That's 80% higher concentration than the biggest tech bubble in history. If these seven stumble, the entire index feels it.
Mag7 = 标普500市值的30–37%。互联网泡沫顶峰时，前7 = 仅19%。这比历史上最大科技泡沫的集中度高80%。如果这七家失误，整个指数都会感受到。
⚠️ Circular Revenue?
⚠️ 循环收入？
Some claim companies buy from each other in a fake circle. Reality: Cross-purchasing is a tiny % of the pie. They all buy from Nvidia, but many are creating cheaper alternatives. Amazon buys ads from Google because it adds value. We, the consumers, are the real source of revenue.
有人声称公司在虚假循环中互相购买。现实：交叉购买在总盘中只占极小比例。它们都从英伟达购买，但许多在制造更便宜的替代品。亚马逊从谷歌买广告是因为它增加价值。我们消费者才是收入和利润的真正来源。
The Bull Case — 90% Cheaper Than Dot-Com
看多理由——比互联网泡沫便宜90%
📊 Valuations: A Fraction of Dot-Com Peaks
📊 估值：仅为互联网泡沫峰值的一小部分
Mag7 trade at ~27× forward earnings. Dot-com average was 52–63×. Cisco alone traded at 200× in 2000. Average dot-com IPO price-to-sales: 48.9×. Today's Mag7: 3–4× price-to-sales. A 90% reduction from dot-com peaks. Mag7 contributed 52% of S&P 500 earnings growth in 2024. $1.98T combined revenue. 25.8% net margins vs 13.4% S&P average.
Mag7以约27倍前瞻盈利交易。互联网泡沫平均52–63倍。仅思科2000年就以200倍交易。互联网泡沫IPO平均市销率：48.9倍。今天Mag7：3–4倍市销率。比互联网泡沫峰值便宜90%。Mag7贡献标普500 2024年盈利增长的52%。$1.98T合计营收。25.8%净利率 vs 13.4%标普均值。
💰 Real Earnings, Not Eyeballs
💰 真实盈利，不是眼球
Fewer than 20% of dot-com companies were profitable at IPO. By Q1 2000: 1 in 5. They burned cash for clicks. Today's Mag7 generate hundreds of billions in free cash flow during the biggest technology transformation most of us will ever see.
互联网泡沫IPO时不到20%盈利。到2000年Q1：5分之一。他们烧钱换点击。今天的Mag7在我们大多数人将见到的最大技术转型中产生数千亿自由现金流。
The Michael Burry Problem
Michael Burry问题
2008 call was legendary. But since then: bearish since 2021, calling it "the greatest speculative bubble of all time." Wrong in 2021. Wrong in 2022. Wrong in 2023. Market kept climbing while he kept shorting.
2008年的判断是传奇。但之后：自2021年看空，称"史上最大投机泡沫"。2021年错。2022年错。2023年错。他持续做空，市场持续上涨。
Latest: $1.1B in put options against AI stocks. Claims hyperscalers understating ~$176B depreciation (2026–2028). Reality: Data center GPUs are upgraded in planned cycles, fully disclosed in SEC filings. Not Enron-style tricks. Standard depreciation audited quarterly. He conveniently deregistered as investment adviser right after his bearish bets failed — no more 13F disclosures required.
最新：$11亿看跌期权做空AI股。声称超大规模企业少报约$176B折旧（2026-2028）。现实：数据中心GPU在计划周期内升级，在SEC文件中充分披露。不是安然式把戏。标准折旧每季度审计。他在看空押注失败后 conveniently 注销了投资顾问注册——不再需要13F披露。
Being right once doesn't make you right forever. When Burry announces a massive short, it moves markets. People doubt their positions. That benefits his bet. At some point, that's influence, not analysis.
对一次不等于永远对。当Burry宣布大规模做空，它影响市场。人们怀疑自己的仓位。这有利于他的押注。到了某个时刻，这是影响力，不是分析。
AI Revenue Is Real
AI营收是真实的
The risk isn't that these companies are worthless. It's that expectations are set so high that any misstep gets punished hard. Corrections are coming — that's part of the process. But a full dot-com crash? The facts don't point that way.
风险不是这些公司一文不值。而是期望设得太高，任何失误都会被严厉惩罚。回调即将来临——这是过程的一部分。但完整的互联网泡沫式崩盘？数据不支持。