Opinion: The AI stock market bubble has arrived. Why am I turning to bet on Bitcoin?
- Core Opinion: The current AI-driven stock market frenzy already exhibits the classic characteristics of a "bubble top." Based on Howard Marks' cycle theory, the author believes market risks far outweigh opportunities and has liquidated tech stocks, turning to buy the dip in Bitcoin, as it offers a wider margin of safety.
- Key Elements:
- Obvious AI Bubble Characteristics: The market exhibits a "price is irrelevant" sentiment, with a massive influx of capital into a single narrative. The stock market's Cyclically Adjusted Price-to-Earnings (CAPE) ratio has broken 40 for the first time, approaching the peak of the internet bubble era.
- Severe Market Disconnect: Despite the Federal Reserve's tightening cycle, the stock market has continued to rise due to AI capital expenditure, fiscal stimulus, and passive fund inflows. However, market breadth is extremely narrow, with AI-related stocks contributing approximately 80% of the gains.
- Risk vs. Opportunity Comparison: The author believes AI stocks are overvalued, with downside risks outweighing upside potential. Conversely, Bitcoin exhibits many characteristics of a "market bottom" (such as negative sentiment and forced selling), offering a safer investment prospect.
- Personal Investment Strategy: The author adopts a "three-bucket" allocation, with 70% designated for long-term accumulation. Currently holding over 80% cash, the author has begun buying spot Bitcoin in batches at an average price of around $59,000.
- Other Market Signals: South Korea's KOSPI was halted twice due to a plunge in AI stocks. Dalio's bubble indicators are near historical highs, and Buffett holds a record $381 billion in cash, indicating selective risk appetite.
Original Author: Investing Beanstock
Original Compilation: Deep Tide TechFlow
Introduction: AI stocks are surging, but this trader is instead liquidating tech stocks and bottom-fishing Bitcoin—he uses Howard Marks' cycle theory to systematically compare the current market, finding that AI already meets almost all the characteristics of a "bubble top." For investors, this article provides a calm framework for cycle positioning, helping you judge whether it's time for greed or fear.
The stock market is experiencing an AI-driven manic bull run, which shouldn't surprise anyone.
If you're not holding any positions, it feels like a fool—because capital expenditure (CAPEX) can only keep rising, and the forward valuations of all these stocks are only getting more insane.
I don't plan to comment on specific stocks or indices; banks and financial media worldwide are already covering them extensively. I'm more interested in figuring out—or at least trying to decipher—what stage of the market we're in. Not just crypto, but the entire financial market.
To that end, I've drawn a lot of inspiration from one of my favorite books: Howard Marks' The Most Important Thing and his work on cycles.
Most people understand cycles as a sequence of events. Most also understand that these events usually follow each other in a conventional order: an upswing is followed by a downswing, and eventually a new upswing. But to fully comprehend cycles, that's not enough. The events in a cycle shouldn't just be seen as happening one after another; more importantly, each one causes the next.
The straight line = the midpoint; the market pendulum is the wavy line oscillating around this midpoint. Together, they form the market cycle, driven by various market forces that cause it to deviate from the midpoint from time to time.
The movement of cyclical phenomena can be easily identified in several stages:
a: Recovering from an excessively suppressed lower extreme or "low" back to the midpoint
b: Continuing past the midpoint to swing to an upper extreme or "high"
c: Reaching the high point
d: Correcting downward from the high back to the midpoint or average
e: Continuing the downward movement past the midpoint toward a new low
f: Reaching the low point
g: Recovering from the low back to the midpoint
h: The cycle repeats
So Where Are We Now?
Is this a bubble? I think it's pretty obvious by now that AI is indeed a bubble. According to Marks, when the sentiment that "price doesn't matter" is strong, that's a hallmark of a bubble.
In a bubble, investors often conclude that you can make money by borrowing (leveraging) to buy a hyped asset. Regardless of your lending rate or funding rate, that asset will surely appreciate faster than that.
"No price is too high" is the ultimate ingredient of a bubble, a fairly clear signal that the market has gone too far.
There are actually conflicting schools of thought, suggesting that markets can trade far above their intrinsic value and still continue to deliver multiples of returns due to mania.
What Should We Do?
Since we're unsure when the bubble will burst, in my view, we have two clear methods for portfolio allocation.
Dollar-Cost Averaging (DCA) (I mean true DCA, no market timing, you just buy little by little in a boring, mechanical way. The more batches you split into, the smoother your final cost basis—that's the whole point of doing it.)
Heavy cash position, but still allowing yourself to participate in the market through tactical/satellite positions, like active trading.
Personally, I prefer the second method. But that's because I actively monitor the market day in and day out, relying on my market experience and intuition to navigate it.
DCA isn't a bad method either. But it does require individuals to truly extend their time horizon. Not 1 or 3 years, but at least 5 to really see some results. Most people DCA for a few weeks, or try to DCA while timing the market, and end up being counterproductive. If you plan to DCA a specific investment, make sure you fully understand the business/industry, then just stick with it in a super boring, repetitive way and continue with your life.
In my 2025 review and reflection post, I mentioned allocating 25% of my positions to passive ETFs, including QQQ, SOXQ, XAR, URA, and UFO. I think a large part of the gains came from QQQ and SOXQ, but I sold them all in May because I believed the market had far exceeded the midpoint.
I also said I was generally bearish on crypto until early 2026 (turned out to be correct), and I managed to hold a lot of cash, patiently deploying some of it into BTC now. The target accumulation range is $50k-$60k, so I've started allocating as I write this.
"Crypto is Dead, Move to AI"
Honestly, my only regret is not allocating more to speculative private market exposure in Anthropic and xAI. I think frontier models still offer the purest AI exposure compared to the public market's "picks and shovels" types, like GPU/semiconductor/storage narratives. Since those are already consensus, I don't think chasing them now offers asymmetric upside. That ship has long sailed. Stocks like MU have nearly 10x'd in a year, and stocks like SNDK are basically moving like memecoins. Upside may still exist, but the downside risk looks worse.
But CAPEX! Yes, this might translate into real value creation in the future, but it's still speculative. Over-speculation, too much money flowing into the same thing—I've seen this movie before.
Is it disappointing to miss a big chunk of the AI bull run? Sure, it stings a bit. But I still have exposure, and it's doubling. I really don't think it's responsible to tell people at today's valuations that AI stocks are truly worth buying unless they really know what they're doing and are in it for the long term (most aren't; they're here for quick money).
Market Sanity Checklist
Now, let's get back to how Marks judges whether we are approaching/nearing a market top:
The economy is growing, economic reports are positive
Corporate earnings are rising and beating expectations
The media only reports good news
Securities markets are strong
Investors are becoming increasingly confident and optimistic
Risk is seen as scarce and mild
Investors believe taking risk is the path to profit
Greed drives behavior
Demand for investment opportunities exceeds supply
Asset prices exceed intrinsic value
Capital markets are wide open, making it easy to raise funds or roll over debt
Defaults are rare
Skepticism is low, confidence is high, meaning risky trades can be made
No one can imagine things going wrong. No favorable development seems impossible
Everyone assumes things will get better forever
Investors ignore the possibility of loss, only worried about missing out
No one can think of a reason to sell, and no one is forced to sell
Buyers outnumber sellers
If the market falls, investors are happy to buy
Prices reach new highs
The media celebrates this exciting event
Investors become euphoric and carefree
Shareholders marvel at their own cleverness: maybe they'll buy more
Those who have been on the sidelines feel regret; therefore, they capitulate and buy
^ This means:
Future returns are low (or negative)
Risk is high
Investors should forget about missed opportunities and only worry about losing money
This is a time for caution!
So, how many of these do you think the current stock market is exhibiting?
On the flip side of the "market top" checklist, the opposite scenario can also occur:
Economic slowdown: reports are negative
Corporate earnings are flat or declining, missing expectations
The media only reports bad news
Markets are weakening
Investors become worried and depressed
Risk is seen as everywhere
Investors believe taking risk is just a way to lose money
Fear dominates investor psychology
Demand for securities is lower than supply
Asset prices are below intrinsic value
Capital markets are tight, making it difficult to issue securities or refinance debt
Defaults surge
Skepticism is high, confidence is low, meaning only safe trades can be made, or none at all
No one believes improvement is possible. No outcome seems too negative to occur
Everyone assumes things will get worse forever
Investors ignore the possibility of missing out, only worried about losing money
No one can think of a reason to buy
Sellers outnumber buyers
"Don't try to catch a falling knife" replaces "buy the dip"
Prices reach new lows
The media focuses on this depressing trend
Investors become depressed and panicked
Shareholders feel foolish and disillusioned. They realize they don't truly understand the reasons behind their investments
Those who didn't buy (or sold) feel vindicated and are praised for their cleverness
Those holding on give up and sell at distressed prices, further exacerbating the downward spiral
^ This means:
Implied future returns are sky-high
Risk is very low
Investors should forget about the risk of losing money, only worried about missing out
This is a time for aggression!
Based on the checklist above, I do think BTC is exhibiting many of these traits (especially the Saylor/MSTR situation). So I genuinely feel that BTC presents a more attractive investment prospect compared to the high-flying AI stocks today.
However, note that the above progressions are simplified. They might not even appear in the same order, nor necessarily in every market cycle. But these behaviors are real, and they are indeed elements that rhyme in markets over decades.
The AI revolution has clearly benefited tech stocks, especially semiconductors over the 1/3/5 year periods.
But investing is never about looking in the rearview mirror (unfortunately, most people do this and draw references from the past). The edge lies where people overlook/dismiss it. We need to look at "what will happen from 1 to 10 years from now," not what the environment is today.
Looking at the chart above, saying you're a crypto investor and should have invested in stocks would look foolish.
Based on the above chart, if you chose stocks, statistically, the likelihood of future underperformance is high.
Also, reading this in 2026 might sound like a joke, but based on past lessons from cycles and an understanding of forward returns/valuation fundamentals, I do believe BTC will outperform stocks in the coming years.
The Most Disconnected Macro Environment Ever
We are also in one of the most disconnected and irrational market environments in history.
Under the new Fed Chair Warsh, interest rates are currently at 3.5-3.75%, and he has publicly taken a hawkish stance. But rates aren't compressing, and the stock market keeps rising, all because AI will cure cancer and everyone will make infinite money forever, right?
The stock market's Shiller Cyclically Adjusted Price-to-Earnings ratio (CAPE) has broken above 40 for the first time, the first such instance since the peak of the dot-com bubble. The US stock market's capitalization is now nearly twice its GDP, with valuations higher than during the 2000 bubble.
Valuation multiple expansion during a tightening cycle is the textbook definition of disconnection.
This disconnect is primarily driven by a three-engine narrative/liquidity machine:
AI CAPEX Super Cycle: Large hyperscalers are spending up to $725 billion in 2026, approaching $1 trillion, now accounting for over 30% of the entire S&P 500.
Late-Cycle Fiscal Stimulus: Corporate and individual tax cuts/tariff rebates boost nominal earnings, even as the Fed tightens policy.
Passive Index Fund Flows: Index funds mechanically pour every dollar of 401(k) pension money into the largest market cap companies, regardless of price. Baby boomers are now forced to buy these hyperscaler stocks at all-time highs, and they keep coming.
Risk Appetite Is Selective
Today, capital is pouring into AI/semiconductors, while everything else, including Bitcoin (the darling of the last cycle), is seeing little growth or bleeding. This isn't a market of universal greed, but one funneling all money into a single narrative (AI and its related verticals).
In 2025, AI-related stocks accounted for about 80% of the entire US stock market's gains. Behind these all-time highs, market breadth is extremely narrow; most stocks aren't even contributing to the rally (unless you're AI-related).
