巴菲特式警告：AI狂熱的冷思考

編輯｜MAX

編者按：下面的文字是一段AI生成的視頻文本，近期在網上流行。這段視頻模仿巴菲特老人家的口氣對AI熱潑了一瓢冷水。我們認為這雖然不是巴菲特本人的話，但其中的一些信息和邏輯值得借鑒。我們把它翻譯過來，供讀者一閱。希望對你有啟發幫助。

英文原文：We're sitting here in late 2025, and I'm watching the biggest tech companies in America spend 400 billion on AI infrastructure in a single year. That's not a typo—400 billion.
中文翻譯：我們現在身處2025年末，我正目睹美國最大的科技公司在一年內投入4000億美元用于AI基礎設施。這不是筆誤——4000億美元。

英文原文：Microsoft. Amazon. Google. Meta. They're all in a race to build data centers that could cover Manhattan. They're buying chips like there's no tomorrow, signing deals worth hundreds of billions with companies that don't even have revenue yet.
中文翻譯：微軟、亞馬遜、谷歌、Meta。它們都在競相建造足以覆蓋曼哈頓的數據中心。它們拼命購買芯片，與那些甚至還沒有營收的公司簽署價值數千億美元的協議。

英文原文：And folks, I'm getting that same sinking feeling I had back in 1999.
中文翻譯：朋友們，我又有了那種不祥的預感，就像1999年時一樣。

英文原文：You see, I'm 95 years old now, and I've lived through enough market cycles to tell you one simple truth: When everyone's running in the same direction, you'd better look behind you to see what's chasing them. Because more often than not, they're running toward a cliff.
中文翻譯：你看，我現在95歲了，經歷了足夠多的市場周期，可以告訴你一個簡單的道理：當所有人都朝著同一個方向奔跑時，你最好看看他們身后有什么在追趕。因為很多時候，他們正跑向懸崖。

英文原文：The AI revolution is real—I won't deny that. But what's happening in the markets right now? That's not investing. That's speculation dressed up in a three-piece suit. And history has a way of punishing speculation, no matter how smart the people involved think they are.
中文翻譯：AI革命是真實的——這點我不否認。但眼下市場正在發生什么？那不是投資。那是穿著三件套西裝的投機。而歷史總有辦法懲罰投機，無論參與者認為自己有多聰明。

英文原文：Let me take you back to 1999. I was sitting in my office in Omaha, reading about internet companies with no profits, no products, and sometimes no plans, trading at valuations that would make your head spin.
中文翻譯：讓我帶你們回到1999年。當時我坐在奧馬哈的辦公室里，讀到關于互聯網公司的報道，它們沒有利潤，沒有產品，有時甚至沒有計劃，卻以高得讓人頭暈的估值進行交易。

英文原文：Pets.com had a sock puppet and a dream. Webvan was going to revolutionize grocery shopping. TheGlobe.com, started by two college kids with $15,000, saw its stock jump 606% on the first day of trading—never mind that they had zero revenue.
中文翻譯：Pets.com只有一個襪子木偶和一個夢想。Webvan打算徹底改變雜貨購物。由兩個大學生用1.5萬美元創立的TheGlobe.com，上市首日股價飆升了606%——盡管他們收入為零。

英文原文：And you know what everyone said about me? "Warren's lost his touch. He doesn't understand the new economy. The old man from Omaha is finished."
中文翻譯：你知道當時大家怎么說我嗎？"巴菲特不行了。他不理解新經濟。那個奧馬哈來的老家伙完蛋了。"

英文原文：I remember one particularly painful dinner party where a young venture capitalist—couldn't have been more than 30—told me I was making the biggest mistake of my career by not investing in internet stocks. He said, and I'll never forget this: "Warren, this time is different. The internet changes everything."
中文翻譯：我記得一次特別尷尬的晚宴，一個年輕的的風險投資家——可能還不到30歲——告訴我，不投資互聯網股票是我職業生涯中最大的錯誤。他說，我永遠忘不了這句話："巴菲特，這次不一樣了?；ヂ摼W改變了一切。"

英文原文：Well, he was half right. The internet did change everything, but it didn't change the laws of economics. By 2002, the Nasdaq had lost 78% of its value. That young VC? I heard he went back to working at his father's car dealership.
中文翻譯：嗯，他只說對了一半。互聯網確實改變了一切，但它沒有改變經濟規律。到2002年，納斯達克指數下跌了78%。那個年輕的風投？我聽說他回去他父親的汽車經銷店工作了。

英文原文：Now fast-forward to today. Nvidia just hit a 5 trillion market cap—5 trillion for a chip company. OpenAI is valued at 500 billion—that's half a trillion—despite posting only 13 billion in projected revenue for 2025. They've committed to spending 300 billion with Oracle over five years, which means they'll need to spend 60 billion a year while bringing in $13 billion. Even my fifth-grade arithmetic teacher could tell you that math doesn't work.
中文翻譯：現在快進到今天。英偉達市值剛剛達到5萬億美元——一家芯片公司值5萬億。OpenAI估值5000億美元——那是半個萬億——盡管其2025年的預計收入僅為130億美元。他們承諾五年內向甲骨文支出3000億美元，這意味著他們每年需要花費600億美元，而年收入只有130億。即使是我五年級的算術老師也能告訴你，這筆賬算不過來。

英文原文：But here's what really concerns me: the circular financing. Nvidia invests $100 billion in OpenAI. OpenAI turns around and spends that money buying Nvidia chips. Oracle builds data centers it hasn't finished yet, for money OpenAI hasn't earned yet. This is exactly—and I mean exactly—what we saw in the late 1990s with telecom companies.
中文翻譯：但真正讓我擔心的是：循環融資。英偉達向OpenAI投資1000億美元。OpenAI轉身就用這筆錢購買英偉達的芯片。甲骨文為OpenAI尚未賺到的錢，建造尚未完成的數據中心。這完全——我的意思是完全——就是我們90年代末在電信公司身上看到的情況。

英文原文：Back then, companies like Global Crossing and Level 3 Communications were laying fiber-optic cable at a breakneck pace. They were lending money to their customers to buy their services, and the customers used that money to buy more capacity. Everyone's stock went up. Everyone felt like a genius.

中文翻譯：那時候，像Global Crossing和Level 3 Communications這樣的公司正在以驚人的速度鋪設光纜。他們借錢給客戶購買他們的服務，客戶又用這些錢購買更多容量。所有人的股票都上漲了。所有人都覺得自己是天才。

英文原文：Four years after the bubble burst, 85% to 95% of that fiber was still sitting dark in the ground, completely unused. Corning, the world's largest optical fiber producer, watched its stock crash from 100 to 1.
中文翻譯：泡沫破裂四年后，85%到95%的光纖仍然埋在地下，完全沒有被使用。世界上最大的光纖生產商康寧，眼睜睜看著其股價從100美元跌到1美元。

英文原文：I didn't invest in any of those companies—not because I'm smart, but because I didn't understand how they were going to make money. My rule has always been simple: If I can't understand it, I won't invest in it. That's what I call my circle of competence, and I've spent 70 years staying inside that circle.
中文翻譯：我沒有投資任何那些公司——不是因為我很聰明，而是因為我不理解他們將如何賺錢。我的原則一直很簡單：如果我不懂，我就不投資。這就是我所說的能力圈，我花了70年時間待在這個圈子里。

英文原文：You might ask, "But Warren, didn't you miss out on the internet boom?" Sure did. Amazon, Google, Microsoft—all of them went on to become giants. And you know what? I'm okay with that. Because for every Amazon that survived, there were hundreds of Pets.coms that didn't. The cemetery of dead dot-com companies is filled with investors who thought they were smarter than everyone else.
中文翻譯：你可能會問，"但是巴菲特，你不是錯過了互聯網繁榮嗎？" 確實錯過了。亞馬遜、谷歌、微軟——它們都成了巨頭。你知道嗎？我覺得沒關系。因為每一個存活下來的亞馬遜背后，都有數百個Pets.com沒能活下來。那些死去的互聯網公司的墓地里，充滿了自以為比別人聰明的投資者。

英文原文：Here's the thing about bubbles that most people don't understand: They're built on a foundation of truth. The internet was revolutionary. E-commerce did change retail. Cloud computing did transform business. But the question isn't whether the technology is real—it's whether the valuations make sense.
中文翻譯：關于泡沫，大多數人不明白的一點是：它們建立在事實的基礎之上。互聯網是具有革命性的。電子商務確實改變了零售業。云計算確實改變了商業。但問題不在于技術是否真實——而在于估值是否合理。

英文原文：In 2000, the Nasdaq reached a price-to-earnings ratio of 200. Read that again—200 times earnings. Today, we're not quite there yet. The S&P 500's Shiller P/E ratio is sitting at about 40, approaching the third-highest level in 154 years of market history.
中文翻譯：2000年，納斯達克的市盈率達到了200倍。再讀一遍——200倍的收益。今天，我們還沒到那個地步。標普500的席勒市盈率約為40倍，正接近154年市場歷史上的第三高水平。

英文原文：But here's what worries me: 80% of the market's gains are concentrated in just a handful of technology stocks—the so-called "Magnificent Seven" (Nvidia, Apple, Microsoft, Amazon, Alphabet, Meta, and Tesla). They account for 75% of S&P 500 returns. That's not a market—that's a house of cards.
中文翻譯：但讓我擔心的是：市場80%的漲幅集中在僅僅幾只科技股——即所謂的"七巨頭"（英偉達、蘋果、微軟、亞馬遜、 Alphabet、Meta和特斯拉）。它們占了標普500回報率的75%。這不是一個市場——這是一個紙牌屋。

英文原文：And the data just keeps getting more concerning. According to recent research, AI-related capital expenditures now account for more than half of U.S. GDP growth in the first half of 2025. Let me repeat that: AI spending is the primary driver of economic growth in America right now—not consumer spending, not business investment across multiple sectors—one technology trend.
中文翻譯：而且數據越來越令人擔憂。根據最近的研究，2025年上半年，與AI相關的資本支出現已占到美國GDP增長的一半以上。讓我重復一遍：AI支出是目前美國經濟增長的主要驅動力——不是消費者支出，不是跨多個行業的企業投資——僅僅是一個技術趨勢。

英文原文：I've seen this level of concentration exactly twice before in my life: once during the "Nifty Fifty" bubble of the early 1970s, when 50 blue-chip stocks traded at absurd valuations and everyone thought they could never go down, and then again in 1999 with tech stocks. Both times ended in tears.
中文翻譯：在我一生中，我只見過兩次這種程度的集中：一次是在1970年代初的"漂亮五十"泡沫時期，當時50只藍籌股以荒謬的估值交易，所有人都認為它們永遠不會下跌；另一次是在1999年的科技股時期。兩次都以眼淚告終。

英文原文：The venture capital numbers are even more alarming. Nearly 64% of all U.S. venture capital in the first half of 2025 went to AI companies. For comparison, internet deals only comprised about 25% of VC investment at the peak of the dot-com boom.
中文翻譯：風險投資的數字更令人震驚。2025年上半年，美國近64%的風險投資流向了AI公司。相比之下，在互聯網泡沫的頂峰時期，互聯網交易僅占風險投資總額的25%左右。

英文原文：We have more than 1,300 AI startups with valuations over $100 million and 498 AI "unicorns" worth more than a billion dollars each. Most of these companies aren't profitable—in fact, 70% of funded AI startups are still losing money—but their valuations keep climbing because everyone's afraid of missing out.
中文翻譯：我們有超過1300家估值超過1億美元的AI初創公司，以及498家各自價值超過10億美元的AI"獨角獸"。這些公司大多不盈利——事實上，70%獲得資金的AI初創公司仍在虧損——但它們的估值不斷攀升，因為每個人都害怕錯過。

英文原文：You know what we call that in Omaha? Gambling.
中文翻譯：你知道我們在奧馬哈管這叫啥嗎？賭博。

英文原文：Charlie Munger, God rest his soul, used to tell me: "Warren, it's waiting that helps you as an investor—and a lot of people just can't stand to wait." Charlie understood something that most people on Wall Street never learn: Making money in the stock market is about patience, not activity.
中文翻譯：查理·芒格，愿他安息，過去常對我說："巴菲特，等待對投資者是有幫助的——而很多人就是無法忍受等待。"查理明白一些華爾街大多數人永遠學不會的道理：在股市賺錢靠的是耐心，而不是頻繁操作。

英文原文：Right now, I look at companies spending 22% to 30% of their revenue on capital expenditures—money they're investing in AI infrastructure. That's three times what utility companies spend, and utilities actually have predictable cash flows. Meta is planning a data center that would require over 2 gigawatts of power—enough electricity to power a small city—just to house 1.3 million Nvidia GPUs. And I have to ask: For what? To train AI models that might be obsolete in two years? To compete in a race where everyone's spending billions to maybe, possibly, hopefully generate returns sometime in the distant future?
中文翻譯：現在，我看到一些公司將其收入的22%到30%用于資本支出——他們投資于AI基礎設施的錢。這是公用事業公司支出的三倍，而公用事業公司實際上擁有可預測的現金流。Meta正計劃建造一個需要超過2吉瓦電力——足以支撐一個小城市的用電——的數據中心，僅僅為了容納130萬個英偉達GPU。我不得不問：為了什么？為了訓練可能在兩年內就過時的AI模型？為了參加一場人人都投入數十億，只為了可能、或許、希望在遙遠的未來產生回報的競賽？

英文原文：This isn't investing. This is speculation on steroids.
中文翻譯：這不是投資。這是打了興奮劑的投機。

英文原文：The uncomfortable truth that nobody wants to hear is this: For all this spending to pay off, AI companies need to generate 2 trillion in annual revenue by 2030. Right now, they're bringing in about 20 billion. That's a 100-fold increase in five years—not a 100% increase, a 100-fold increase.
中文翻譯：沒人想聽的不舒服的真相是：要讓所有這些支出獲得回報，AI公司需要在2030年前實現2萬億美元的年收入。目前，它們的收入約為200億美元。這是在五年內增長100倍——不是增長100%，而是增長100倍。

英文原文：I've been in this business since 1951. I've never—and I mean never—seen an industry scale that fast while maintaining profitability. It defies the laws of business gravity.
中文翻譯：我自1951年就從事這個行業。我從未——我意思是從來沒有——見過一個行業在保持盈利能力的同時，能以如此快的速度擴張。這違背了商業引力定律。

英文原文：Let me tell you something about numbers: They're stubborn things. You can dress them up, put lipstick on them, call them "adjusted EBITDA" or "AI-adjusted revenue," but at the end of the day, a business either makes money or it doesn't. And right now, the numbers in AI don't add up.
中文翻譯：讓我告訴你關于數字的一件事：它們很固執。你可以打扮它們，給它們涂上口紅，稱它們為"調整后息稅折舊攤銷前利潤"或"AI調整后收入"，但歸根結底，一個企業要么賺錢，要么不賺錢。而現在，AI領域的數字根本對不上。

英文原文：Microsoft, Amazon, Google, and Meta are projected to spend a combined $364 billion in their 2025 fiscal years. These aren't startup companies gambling with venture capital—these are the largest, most profitable corporations in the history of capitalism, and they're spending money like drunken sailors on shore leave.
中文翻譯：微軟、亞馬遜、谷歌和Meta預計在2025財年將總共支出3640億美元。這些不是用風險資本賭博的初創公司——這些是資本主義歷史上最大、最盈利的公司，而它們花錢卻像上岸休假的喝醉的水手。

英文原文：Microsoft just raised its capital expenditure guidance to over 120 billion for 2025. Amazon is spending 100 billion. Google bumped its forecast from 75 billion to 85 billion. Meta's at 70 to 72 billion and announced they expect "similarly significant CapEx dollar growth in 2026."
中文翻譯：微軟剛剛將其2025年的資本支出指引提高到超過1200億美元。亞馬遜計劃支出1000億美元。谷歌將其預測從750億美元提高到850億美元。Meta在700億到720億美元之間，并宣布他們預計"2026年資本支出金額將有類似的大幅增長"。

英文原文：Now, I want you to understand what these numbers mean in real terms. Microsoft's $120 billion in capital spending is seven times what they spent just five years ago. Seven times. That's not growth—that's a complete transformation of their business model. And they're betting the farm that it works.
中文翻譯：現在，我想讓你們理解這些數字的實際意義。微軟1200億美元的資本支出是五年前支出的七倍。七倍。這不是增長——這是他們商業模式的徹底轉變。他們正在押上全部身家指望它能成功。

英文原文：But here's where it gets interesting: Despite all this spending, AI-related revenue remains a tiny fraction of their total business. Microsoft reported that their AI services had reached a 13 billion annual run rate. That sounds impressive—until you realize they're spending 120 billion to generate 13 billion in revenue. That's like spending $9 to make $1.
中文翻譯：但有趣的地方在這里：盡管有所有這些支出，AI相關收入仍然只占他們總業務的很小一部分。微軟報告稱，其AI服務已達到130億美元的年化運行率。這聽起來令人印象深刻——直到你意識到他們花費1200億美元來產生130億美元的收入。這就像花9美元去賺1美元。

英文原文：But Warren, you might say, "This is about the future. They're investing for long-term gains." And you're right—it is about the future. But there's investing for the future, and then there's speculation. The difference is simple: Can you reasonably predict the return on that investment?
中文翻譯：但是你可能會說，"巴菲特，這是為了未來。他們是在為長期收益進行投資。" 你說得對——這確實是為了未來。但有為未來的投資，也有投機。區別很簡單：你能合理地預測那項投資的回報嗎？

英文原文：In the dot-com era, companies like Global Crossing spent 15 billion building a worldwide fiber-optic network. The logic was impeccable: Internet traffic was doubling every 100 days, so we'd need massive capacity. Except they overbuilt by 99%. When the music stopped, that 15 billion in investment was worth pennies on the dollar.
中文翻譯：在互聯網時代，像Global Crossing這樣的公司花費了150億美元建設全球光纖網絡。邏輯是無懈可擊的：互聯網流量每100天翻一番，所以我們需要巨大的容量。只是他們過度建設了99%。當音樂停止時，那150億美元的投資只值原價的幾分之一。

英文原文：Today's AI infrastructure build-out is happening even faster and at an even larger scale. The parallel isn't just similar—it's practically identical.
中文翻譯：今天AI基礎設施的建設速度更快，規模更大。這種相似之處不僅僅是類似——簡直是如出一轍。

英文原文：And here's something that should terrify anyone who understands market history: This spending is being funded increasingly by debt. Meta just arranged $27 billion in off-balance-sheet financing. These companies are taking on leverage at the same time they're making massive, uncertain bets on technology that might not pay off for years, if ever.
中文翻譯：還有一點應該讓任何了解市場歷史的人感到恐懼：這些支出越來越多地由債務提供資金。Meta剛剛安排了270億美元的表外融資。這些公司在承擔杠桿的同時，正在對一項可能多年后——如果真有回報的話——才能見效的技術進行大規模、不確定的押注。

英文原文：I've lived through enough credit cycles to know that debt magnifies everything—both gains and losses. When things are going well, debt makes you rich faster. But when things turn, debt is what kills you.
中文翻譯：我經歷了足夠多的信貸周期，知道債務會放大一切——無論是收益還是損失。當事情進展順利時，債務讓你更快致富。但當情況轉變時，債務會要了你的命。

英文原文：Remember, in my early days, I made the mistake of buying Berkshire Hathaway itself because I was angry at management. It was a "cigar butt" stock—cheap enough that I thought I could get one more puff out of it. That investment, which I made out of emotion rather than reason, probably cost me $200 billion in opportunity cost over the years.
中文翻譯：記住，在我早期，我犯過錯誤，因為對管理層不滿而買入了伯克希爾·哈撒韋公司本身。那是一支"煙蒂"股——足夠便宜，我以為我能再吸最后一口。那項出于情感而非理性的投資，這些年來可能讓我損失了2000億美元的機會成本。

（文本來源視頻號“陳佳9527”《任何投資人都要看一百遍》）

免責聲明：本文基于公司法定披露內容和已公開的資料信息整理，文章不構成投資建議僅供參考。

睿藍財訊出品

文章僅供參考 市場有風險 投資需謹慎

來源：藍籌企業評論