저자: TomLee · 외국인
요약 보기
Tom Lee는 AI는 버블이 아니라 이제 막 시작된 인프라 투자 물결이라고 주장한다. 빅테크의 약 3조달러 규모 AI 지출 약정은 아직 집행되지 않은 것이라 대차대조표에 안 잡히는 것이 정상이며, 닷컴 시기 IRU 같은 순환 거래와 달리 지금의 현금 부유한 기업들이 핵심 사업에 재투자하는 것이라고 설명한다. AI 도입률은 약 7% 수준이라 성장 여지가 크고 제본스 역설로 AI 사용은 더 늘어난다고 본다. 국가부채 40조달러는 구조적 우려지만 채권시장이 수용하고 있어 당장은 괜찮다는 입장이다. S&P500 연말 7,700, 8월 초 8,000 터치 전망을 언급했다.
다듬은 글
Hello everyone, welcome to the ML Investment Channel. I am Tom Lee, co-founder of Fundstrat Global Advisors . I said yesterday that AI is by no means a bubble; it is an epic wave of infrastructure technology that you only get to see once in a lifetime, and what's more, it has only just begun. Let's look at this together: I was a tech analyst during the dot-com fiber boom of the 90s and, uh, the people investing capital at that time were not of the same ilk and caliber as the MAG 7.
You know, these were companies that were digging up railroad lines and, uh, doing those IRUs, you know, which was actually what an IRU? It was a, uh, revenue swap between fiber companies. So you could create hundreds of billions of dollars...and we have companies, you know, with some of the highest profit margins on capital in history, as these guys have delivered trillions of shareholder returns that are now directing their investment on building an AI network.
So, I have a lot more confidence that these are high-level, well-run companies investing, but these are eye-popping numbers, but the real... Why not do it on balance sheet? Well, one, uh, they could do it on balance sheet, but if they did, they would be taking up all the capital of the world and all the risk and therefore actually make it harder for any...I would argue that that would make it harder for democra-ounting...question.
w.e. we were calling...I did get a...I got so I need to ask...off-balance commitments, because future commitments that these are not noting anything, they're not on the balance sheet because they don't flow through the balance sheet until the building actually begins. So, in other words, are they this is just a different way to look at if they say we're going to spend whatever amount in 2027.
Can you explain to me exactly what these commitments are, and why they're not on the balance sheet, and when they will be? Yeah, um, well, I think maybe a good place to start is, I think...J, ete picture. You know, the gross obligations of the financial system, it's multiple times the underlying assets.
Always, or just today? Always. In fact, that's why Warren Buffet used to call credit derivatives, you know, the weapons of mass destruction. Are these credit derivatives that we're talking about, though, or these are just future spend commitments? It's the...not that, because if you did gross exposure of swaps, options, like looking, any options contracts are multiples of cash underlying.
So, if someone says, "Oh, there's a hidden off-balance-sheet risk that retail investors have 20 times the size of the stock market in bets," we'd be like, "Well, there's the offsets." I , I would say, when we look at these numbers, uh, it's a distorted view of the actual. Let me just press one more.
In other words, do you think that the spending is representing multiple possibilities of spend that's only going to manifest in one way? I think to me, um, none of these contracts like are going to lead to criminal liability . Like, in other words, like a company can decide to cut spending in the future and the contract should be...so, like, the 3 trillion isn't like...
you know people have to like sign over their kidneys to 40 trillion in the national debt last night. What is your reaction to or take on that? I mean, it's pretty shocking that we're in a booming economy and the deficit growing. Um, so misses are expensive. Yeah. Interest. Yes. Yeah. It seems like there's a lot of like outflows.
I, I agree with the folks who think that this is a structural concern, but the bond market today is signaling that it's completely okay with it. So, I think as long as it's okay with it, the stock market is fine. So that that little blip in the 30 year we saw, those like oh the deficit suddenly an issue that we saw last week with that that sale.
You don't think that that was really a big move or that that's going to continue? Yeah. I mean it is it is a symbolic rate but as you know maybe more companies are dependent on the spread to the rate. So that's the more important to watch but again like I don't think it makes any sense to not be sound as a nation.
So... Let's start with the most alarming number. The Wall Street Journal published a report a while ago, titled "Big Tech's spending on AI is 3 trillion dollars higher than what is seen on the surface." The host was particularly sharp: is this money coming from private credit funds, subsidiary companies, or third-party bond issuances?
And who is ultimately on the hook? Tom Lee's response is the part of the whole conversation that I think is the most valuable. He made three main points. First, this money is not hidden at all; it is money that hasn't been spent yet. Accounting standards dictate that commitments are only recorded on the balance sheet when construction actually begins.
If you sign a contract for a data center that doesn't start construction until 2027, it remains invisible on the books—not because you want to hide it, but because it hasn't happened yet. It's just like promising to treat a friend to dinner next year; that meal doesn't exist in this year's ledger—it is merely a commitment.
Second, questioning this by pointing to off-balance-sheet financing just shows that you don't understand how the financial system works. Tom Lee said that if you calculate the total exposure of the entire financial system , it is always many times that of the underlying assets, and this is not a problem specific to today, but something that is always the case.
Buffett calls credit derivatives " weapons of mass financial destruction" because their notional value is dozens of times that of the financial market, yet no one considers them a hidden risk , as everyone knows they are hedged and offset. Options, futures, and forwards are all like this; a huge notional amount does not equate to huge actual risk.
Thirdly, none of the clauses in these contracts will send you to jail, nor will they bankrupt the company— what does that mean? It means these are three-to-five-year procurement commitments; companies can absolutely cut spending in the future, and at most , they’d pay a breach of contract fee —it’s not like they signed a " life-or-death" pact.
Therefore, the $ 3 trillion is not $ 3 trillion that must be spent, but rather a sum of possibilities; in the end, only one path will truly be realized. Someone might ask, even if it's not Enron, why does this still sound like a bubble? Wasn't the 2000 internet bubble created by burning cash exactly like this?
Tom Lee answers this question with great confidence because he was an eyewitness to that bubble back then. He was a tech analyst in the 90s and saw firsthand how the fiber-optic bubble grew. He says the players back then and those of today are not the same at all. Who were the people back then? They were the ones laying railroads and signing IRUs —what is an IRU?
To put it simply, it was fiber-optic companies buying capacity from each other, selling from the left hand to the right, and the right hand back to the left, creating tens of billions in contract revenue out of thin air. No real money was earned, but the financial statements looked great; it was a pure capital game.
Now let’s look at today; those throwing money around are Microsoft, Google, Amazon, and Meta. They are among the most profitable companies in the history of the business world; they have moats, the highest profit margins in the market, and hundreds of billions to trillions in shareholder returns. They aren't raising capital because they lack money; they have so much money they don't know where to spend it , so they decide to plow it back into their core businesses.
There is a very interesting point here; another guest in the conversation added a remark that I think is particularly classic. What did these companies do when they made money in the past? They bought back their own stock, right? But now, they say, "I'm not buying back stock; I'm going to invest in what I do best." They were never hedge funds to begin with; this is what they were always meant to do.
If a company reinvests its free cash flow into its core business, that is the best news you could possibly hear when investing in growth stocks. Tom Lee says this is like Bezos ; Bezos once said that to judge if a company is a good one, look at whether it dares to reinvest its profits into building a new moat.
Today, these giants are spending real money to dig a new moat for themselves centered around AI. Of course, Tom Lee also admits these figures do look quite intimidating, and he says not all railroad stocks made money; if you bought random railroad stocks in the 1860s, you were likely to lose money, but that doesn't change the fact that back then, as long as you were in the railroad industry, you were part of the trend.
Now, let's discuss the question we care about most: the money has been spent, but is the tech actually useful? The host asked a practical question: many stocks are priced based on outrageously large revenue expectations two or three years down the line—can these expectations actually be realized ? They provided data showing that while the vast majority of companies have purchased AI accounts, only about 30% say they have seen an increase in labor productivity, and actual full-scale implementation is even lower; this means this revolution is only 7% complete, implying that 93%of the growth is still ahead.
It’s as if the railroad had only reached Chicago and you were already asking if the railroad industry had reached its end. Tom Lee’s exact words were that we are moving from infrastructure to implementation, from proof of concept to rapid adoption; translated into plain English, that means everyone was buying land and shovels to dig holes a few years ago, and now is the time to plant things, and planting is the stage where you actually make money.
He also mentioned an economic concept called the Jevons Paradox. In the 19th century , an economist named Jevons discovered that the more efficient steam engines became, the more coal was actually burned, because increased efficiency lowered costs, leading more people to use it, causing total consumption to skyrocket.
AI is actually the same; the cheaper and more useful AI becomes, the more companies will use it, and total AI spending will only snowball rather than shrink. That is why he dares to say don't be afraid of the horror stories about AI replacing humans and causing unemployment, though the host followed up by asking, doesn't productivity growth just equal layoffs?
Tom Lee says yes, but every technological revolution has been like this. Back when horse carriage drivers saw the Ford Model T, they were furious ; yet later, they all went to work in factories, earning even more than before. Carriage drivers became car drivers, and workers became programmers . Structural transformation brings growing pains, but the overall trend is upward.
After that, the conversation raised a big question: US national debt has topped 40 trillion dollars, are you panicked? Tom Lee’s answer was very interesting. He said that with the deficit still growing during such an economic boom, it is indeed scary and represents a structural problem. But notice this "but," the bond market is currently voting with real money, saying, "I'm perfectly okay with this." The 30-year Treasury yield did jump last week, but that was more symbolic; what you really need to watch is the 10 -year, because corporate financing costs are pegged to it.
As long as the bond market doesn't collapse, the stock market will be just fine. He also mentioned that the 10-year yield fell below 4%when the Iran war broke out, yet now that oil prices have retreated, it’s actually near 4.7%, which makes no logical sense. But that is just how the market functions.
They say that for a country not to maintain fiscal soundness is unjustifiable, and it is a real problem, but it isn’t today’s problem. The takeaway for ordinary investors from this discussion is not to let macro figures like 40 trillion scare you into shorting, nor let numbers like 3 trillion scare you into missing out.
You must distinguish between signals that can change your investment decisions and those that are just noise. Moomoo, a brand under a US-listed company, has already become the top choice get up to 100 dollars in stocks, and idle cash can earn up to 8.1%returns. If interested, click the link below the video to experience it.
Wrapping up, the core of their conversation can be summed up in three points: first, 3 trillion is not a hidden bomb, but a commitment yet to be fulfilled; off-balance-sheet financing is not Enron-style fraud, but a normal promise under accounting standards, and a standard way the financial system operates.
The real deciding factor isn't whether the money has been spent, but whether labor productivity has improved after spending it. If AI cannot change productivity, then it is just a cute toy, and no amount of money will make it useful. These are the exact words of Tom Lee, and I think it is the clearest statement among all the bullish arguments.
The investment entities this time are completely different from those in the year 2000. Back then, it was shell companies burning other people's money; today, it is the most profitable companies in the world betting their own earnings on their core businesses after careful deliberation by their boards—that is a fundamental difference.
Third, the wave has only reached 7%completion, and the Metcalfe’s Law guarantees that this wave will only get bigger. From infrastructure to practical applications, from proof-of-concept to rapid adoption, the true golden era has just begun. So, the real question has never been whether AI is a bubble, but rather: in this wave, do you want to be a passenger or the one building the road?
Tom Lee predicted the S&P 500 would hit 7,700 by year-end, and in early August, he even called for it to touch 8,000; he is a perpetual bull. You can certainly question his specific price targets, but there is one thing I believe: when a country's capital expenditure reaches 2.5%of its GDP. When the penetration rate of the entire industry is only 7%.
When the smartest money is still pouring in at an accelerating rate, this isn't the end of a bubble; it looks more like the curtain has just been raised. next time!
Comments 0