Welcome back to the Deep Dive. We're digging into 2025 today, a year that, uh, on paper looked fantastic. S&P 500 up 16%, Nasdaq hue up 20%, a roaring bull market, right? But if you actually lived through it, you know it was. Well, this is a rollercoaster. Our mission today is to unpack
that. We're going to get into the real story behind the Fed's sudden pivot, figure out how much of this growth was just the AI engine and what hidden risks started bubbling up for 2026.
That's the core tension right there. The massive gains versus that extreme volatility. That's the signal we need to decode. We're looking at the surprising policy shifts, especially from the Fed and the, you know, the structural risks that your source material lays out so clearly. And of course, before we jump in, just a quick reminder, everything we talk about here is based on public
這就是核心矛盾所在。巨大的收益與極端的波動性。
00:45
information. It's for educational purposes. And it is definitely not investment advice. But on that note, if you are trying to track the markets overall temperature, we do encourage you to search and download the iOS app buffet indicator from the Apple App Store. It's pretty useful tool.
這僅限於公開資訊。這僅供教育目的，絕對不是投資建議。但在此同時，如果您想追蹤市場的整體溫度，我們鼓勵您在 Apple App Store 搜尋並下載 iOS 應用程式 Buffet 指標，這是一個相當有用的工具。
00:58
Okay, let's unpack this. The biggest shock of the year, I think, has to be the Fed.
好，我們來分析一下。我認為今年最大的衝擊非聯準會莫屬。
01:02
The market went into the year expecting what maybe one rate cut, maybe one. And instead, we got this incredibly sharp contrast. The Fed cut rates three times in just six months, September, October and December. It was a sustained, aggressive pivot that just forced
everyone to rethink the entire game plan. Right. And the Fed is always trying to walk this tightrope, isn't it? Maximum employment on one side and stable inflation on the other.
That's their dual mandate. It is. And for these cuts to be so aggressive, one of those goals must have been flashing some serious warning signs. And it wasn't inflation.
No, that's the critical part. Inflation was still sticky. It was floating between, say, 2.5 and 3%, which is still above their 2% target. But the Fed seemed okay to tolerate it.
These three cuts are almost entirely driven by this wolf is coming story around employment, a preemptive strike. You see it with the September cut. They called it a preventative cut because unemployment had already ticked up to 4.4%. The message was, we're not in a catastrophe now,
but we're trying to prevent one. And then the October decision, that was just extraordinary.
但我們正在試圖預防它。而十月的決定更是非比尋常。
02:07
It happened during a government shutdown, which means the main data they rely on, the big payroll reports from the Bureau of Labor Statistics, it just wasn't there.
這發生在政府關門期間，這意味著他們依賴的主要數據，即勞工統計局發布的大型薪資報告，根本沒有。
02:15
Exactly. The source is called it a blind flight.
沒錯。消息來源稱之為盲飛。
02:17
A blind flight. So they were making a huge decision with no data.
盲飛。所以他們在沒有數據的情況下做出了巨大的決定。
02:22
No primary data. And that immediately just fueled speculation that the weakness was already so obvious, or maybe the political pressure was so intense that they had to act anyway. It turned policy into a real gamble. So if they couldn't see the big government numbers, what were they looking at? What were those other signals they were picking up on during this blind flight?
Well, they were relying on alternatives and revisions. So first you have the ADP data, the small payroll numbers from the private sector. That was showing weakness. But the more alarming thing was a historical revision. The Bureau of Labor Statistics had to issue this significant
historical downward revision for job growth in the months before all this.
歷史性向下修正。
03:01
So they were basically admitting that the job market had been much weaker than the headlines we're telling us all along.
所以他們基本上承認就業市場一直比媒體報導的要弱得多。
03:07
Precisely. It confirmed that bad feeling a lot of consumers had. And we also know they look at non-traditional real-time data, things like job postings on recruitment websites, salary trends, the number of open positions. All of those were likely flashing red.
I think that's a fair way to put it. It has the structure of a private institution, but its power is fundamentally public. The highest body, the Federal Reserve Board, has seven members, all appointed by the president.
But they have those long 14-year terms to try and insulate them.
但他們有長達 14 年的任期，試圖讓他們不受干擾。
03:47
They do. But the real power is with the Federal Open Market Committee, the FOMC, 12 voting members. And the chair, Jerome Powell, he only gets one of those 12 votes.
It's just one. And what 2025 really showed wasn't just healthy debate inside the Fed.
僅僅一票。而 2025 年真正顯示的，不僅僅是聯準會內部的健康辯論。
04:03
It was the rise of what the sources call strong division.
而是所謂的「強烈分歧」的興起。
04:06
Deep political friction made worse by all the external attacks from figures like Trump.
嚴重的政治摩擦，因川普等人物的外部攻擊而加劇。
04:11
And you had these factions formed inside, right? The doves, the hawks.
而你們內部形成了派系，對吧？鴿派、鷹派。
04:15
Exactly. You had the extreme doves who were all for cutting, the extreme hawks who wanted to hold or even hike, and then a swing middle.
沒錯。有極端的鴿派主張降息，極端的鷹派希望維持利率甚至升息，然後是中間搖擺的群體。
04:22
And the thing is, you can always find data to support your narrative.
問題是，你總能找到數據來支持你的論點。
04:25
The doves would point to weak employment. The hawks would point to inflation still being above two percent. And the doves won out, which puts Powell in a really tough spot as his term ends in June, 2026. It's the classic political trap. He's either going to be blamed for a recession if he holds rates too long or for runaway inflation if he gets too much. The fact that
there's such strong division, it really suggests he's lost a lot of that soft power.
存在如此強烈的分歧，確實表明他已經失去了很多軟實力。
04:48
A chair's real influence isn't his one vote. It's his ability to build consensus.
主席真正的影響力不在於他的一票。而在於他建立共識的能力。
04:52
That's clearly eroding. And all that instability at the Fed just spilled right into the markets.
這顯然正在侵蝕。而聯準會的動盪也直接影響了市場。
04:58
Let's shift there to volatility. It feels like the markets reaction to political news showed the Fed had kind of lost control of the narrative. Absolutely. The year was defined by it.
I mean, things were calm until April 2nd when President Trump announced these widespread tariff plans. Yeah. And that just triggered instant widespread panic. How bad did it get?
In just a few weeks, from mid February to early April, the market plunged 19%.
在短短幾週內，從 2 月中旬到 4 月初，市場下跌了 19%。
05:21
It's a huge investment pitfall. And it was global. The Hang Seng Tech index in Hong Kong dropped over 17% in a single day. Wow. Yeah. And investors who panicked and sold got hit twice. They took the massive loss and then they missed the explosive rally that came right after. And that rebound
was just as fast, wasn't it? Almost instantaneous. The moment Trump backed off on the most extreme China tariffs and negotiations started, the market just exploded. The D.O.W. jumped almost 8% nasty Q over 12%. In one day, it shows you the key financial data of 2025. Wasn't just corporate
earnings. It was political headlines, which brings us to maybe the biggest theme of the year, this case shaped economy. GDP grew over 2%, which sounds healthy. But when you look under the hood, it's incredibly narrow. It's shocking when you break it down. Yeah. So 2% US GDP is roughly 500
收益。而是政治頭條新聞，這讓我們來談談今年也許最大的主題，這個塑造經濟的案例。國內生產總值（GDP）增長了 2% 以上，聽起來很健康。但當你深入研究時，你會發現它非常狹窄。分解開來令人震驚。是的。所以 2% 的美國 GDP 大約是 500
06:11
billion dollars. The total AI related capital spending all the hyperscalers building data centers buying chips was over 400 billion. So basically the entire US economy outside of the AI ecosystem was essentially flat. It barely grew at all. That's the stagnation. That's the bottom leg of the K
十億美元。所有超大規模雲端服務供應商用於建造數據中心、購買晶片的與 AI 相關的總資本支出超過 4000 億美元。所以基本上，美國經濟除了 AI 生態系統之外，基本上是持平的。幾乎沒有成長。這就是停滯。這就是 K 型曲線的底部。
06:28
while the AI winners are shooting up on the top leg. Exactly. That's why the S&P 500, which is weighted to those big winners, was up 16%. But 37% of the stocks in it were actually down for the year. If you look at the equoid S&P index, it was only up 10%. That gap tells the
而 AI 的贏家則在頂部飆升。沒錯。這就是為什麼市值加權的標普 500 指數，由於這些大型贏家，上漲了 16%。但其中 37% 的股票實際上在當年下跌。如果你看等權重的標普指數，它只上漲了 10%。這個差距說明了
06:43
whole story. And while we're on the K shape, we have to touch on the affordability crisis.
整個故事。既然談到了 K 型曲線，我們就必須談談可負擔性危機。
06:48
Housing especially, it just keeps getting worse for that bottom leg. The sources are very clear on this. It's a supply problem, not a demand problem. The real issues are local, restrictive zoning laws, complicated permitting these crazy impact fees that can add 10 to 15% to the cost
of a new home. The federal government really isn't equipped to solve that. Okay, so let's turn to that engine, the top arm of the K AI. The market is betting everything on this.
But what are the risks that could actually derail this whole cycle?
但有哪些風險可能會真正破壞整個週期？
07:17
The demand is real. No question. I mean, NVIDIA's revenue grew 60%. But there are two huge constraints emerging. The first one is structural. It's the power supply. The actual electricity.
The actual electricity. US electrical demand growth has tripled to 3% a year. That requires 125 terawatt hours of new power. To put that in perspective, that's like added the entire state of California's consumption to the grid. And you can't just build that overnight? No, it takes two to three years, best case. And delays are already happening.
If the hyperscalers can't get the power, they have to slow down their chipboarders, and that whole growth story starts to reverse. Okay, so that's the structural risk. What about the second one? You said it was intellectual. Right. This one is more fundamental. The whole boom is based on the assumption that large language models, LLMs have to get bigger and bigger to
reach artificial general intelligence AGI. More scale means more chips.
達到通用人工智能（AGI）的假設。規模越大意味著晶片越多。
08:09
But what if that's not true? What if we're hitting a point of diminishing returns?
但如果不是這樣呢？如果我們達到了邊際報酬遞減的點呢？
08:12
That's the minority opinion that's starting to grow. People pointed to the release of chat GPD five in August and said, you know, it's not really that much better than version four.
That's a huge risk. But there's another quieter threat brewing in private credit.
這是一個巨大的風險。但還有另一個更隱藏的威脅正在私人信貸領域醞釀。
08:34
It's ballooned from what two trillion to three trillion dollars since 2020?
自 2020 年以來，它已經從 2 兆美元膨脹到 3 兆美元？
08:38
And the keyword there is quiet. The biggest risk is its opacity. Back in 2008, mortgage data was public. It was complex, but you could see it. With private credit, there's no central database. If it starts to collapse, it's going to come as a complete surprise to the public. We saw some warning signs, though. We did two big fraud cases in 2025.
Tri-color and first brands, they involved misrepresenting collateral. And they just signal that there are these systemic weaknesses, this willingness to cheat that we get exposed in a big way when the next recession finally hits. Let's zoom out to the global picture for a minute.
Tri-color 和 First Brands，它們涉及對抵押品進行虛假陳述。它們只是表明存在這些系統性弱點，這種欺騙的意願，當下一場衰退最終來臨時，我們將會大規模暴露。讓我們稍微放大一下，看看全球圖景。
09:13
With the Fed cutting rates, what's happening in China? Well, the Fed cuts are narrowing the interest rate gap between the US and China. That takes pressure off their currency, the R&B.
And that gives the People's Bank of China a lot more room to maneuver to use their own monetary policy to deal with their domestic challenges. And one of those challenges is driving them into this low interest rate era. Why are their rates falling so consistently? It's a mix of things.
Cyclically, they have low inflation and weak consumption. So if they don't cut nominal rates, their real interest rates stay high, which just encourages more saving instead of spending.
But it's also structural. They need cheap long-term money for what they call high quality development, EVs, AI, chip manufacturing. Those are industries that need huge upfront investment and have long payoff cycles. That plus things like demographic changes and slowing urbanization,
it all just naturally pushes long-term rates down. And we have to mention two of the big investment landmines from 2025. Bitcoin hitting $126,000 and then crashing. Oh, that was a textbook lesson.
The surge was fueled by massive leverage. When that tariff shock hit the market, it triggered a cascade of margin calls, forced liquidations, a total liquidity vacuum. It proved once and for all that it behaves like a high-risk tech stock, not some kind of safe haven. And gold had a big
這波飆升是由龐大的槓桿所推動。當關稅衝擊 hit 市場時，引發了連鎖的保證金追繳、強制平倉，以及全面的流動性真空。這一次又一次地證明，它表現得像高風險科技股，而不是某種避險資產。黃金也出現了大幅
10:32
surprising drop too. It did. Gold fell 6% in one day in October, the biggest drop in years.
出人意料的下跌。確實如此。黃金在十月一日下跌了 6%，是多年來最大單日跌幅。
10:38
That was driven by a sudden hope for less geopolitical conflict plus a stronger dollar.
這是由於對地緣政治衝突減少的突然希望，加上美元走強所驅動。
10:42
It really showed that gold's value was tied more to the narrative of global tension than anything else. So what does this all mean? As we look ahead to 2026, what are the big takeaways?
I think there are two core risks that define the path forward. First is stubborn inflation.
我認為有兩個核心風險定義了前進的道路。首先是頑固的通貨膨脹。
10:56
Even with the Fed cutting, it might just stick around near 3%. That means we could easily be looking at a mild stagflation environment, low growth, but with stubbornly high inflation.
And second is that threat from private credit. If the Fed keeps policy easy, it just encourages more and more leverage in those opaque markets. We know the weaknesses are there.
When the next recession hits, those weaknesses will be exposed, and it could trigger a pretty significant credit crisis. That sounds like a tough road ahead, but it also kind of confirms that every year is hard paradox. Markets somehow navigable at these things. I guess our focus
should be on understanding these mechanisms you've laid out, not trying to make perfect predictions. Exactly. And it should lead us to ask a more provocative question for the future.
Given all the political pressure we're seeing on central banks around the world, should we just expect that their independence is going to keep eroding, making all our future economic cycles even more volatile and tied to election cycles? That is a crucial question to keep in mind. Thank you for joining us on this steep dive. And just a reminder for everyone
listening, this content is for educational purposes only. It's not. Full in-depth analysis, you're welcome to join our Deep Value Club membership or subscribe to us on Apple and Spotify podcasts. The channel is Deep Value Investing. We'll see you next time.
收聽，本內容僅供教育目的。它不是。深入的分析，歡迎您加入我們的 Deep Value Club 會員，或在 Apple 和 Spotify 播客上訂閱我們。頻道是 Deep Value Investing。下次再見。
影片強調，超大規模雲端服務供應商在 AI 相關的資本支出超過 4000 億美元，幾乎相當於美國總 GDP 增長約 5000 億美元，表明它帶動了大部分的增長。
3WhatarethetwomainconstraintsemergingfortheAIgrowthstorymentionedinthevideo?影片中提到的 AI 成長故事出現的兩個主要限制因素是什麼？WhatarethetwomainconstraintsemergingfortheAIgrowthstorymentionedinthevideo?
6Whatdoesthe 'K-shapedrecovery' or 'K-shapedeconomy' refertointhecontextofthevideo?影片中提到的「K型復甦」或「K型經濟」指的是什麼？Whatdoesthe 'K-shapedrecovery' or 'K-shapedeconomy' refertointhecontextofthevideo?
10WhatisakeystructuralrisktotheAIgrowthstorymentionedinthevideo?影片中提到的 AI 成長故事的一個關鍵結構性風險是什麼？WhatisakeystructuralrisktotheAIgrowthstorymentionedinthevideo?
