Author: Campbell , Macro Analyst

Compiled by: Yuliya, PANews

Editor's Note: Recently, the US memory chip sector has become the main theme of the technology market, with companies like Micron Technology, SK Hynix, and SanDisk continuing to see significant stock price increases. Meanwhile, the debate over whether AI has entered a bubble phase has intensified again. Opinions vary widely: Dan Niles, a renowned chip analyst from the dot-com bubble era, believes that current AI development is closer to the mid-sprint of internet infrastructure construction in 1997 than the tail end of the bubble in 1999. He points out that the rise of AI agents is driving a surge in computing power demand, and while chip stocks are currently overvalued, they still have long-term potential. Hedge fund legend Paul Tudor Jones also predicts that the AI bull market has likely completed about 50% to 60% of its run and may continue for another one to two years. In contrast, Michael Burry, the real-life inspiration for the protagonist of the movie "The Big Short," warns that the current market is highly similar to the eve of the dot-com bubble burst in 2000.

Amidst the current fervor and anxiety, and with industry leaders offering conflicting opinions, if a bubble truly exists, how should we respond? This article, drawing on the author's own experience, shares a hard-core, practical guide on "how to short a bubble." Below is the original article:

Frankly, I don't know if we're in a bubble right now, I'm not even sure if it's a question that can be known. My understanding is similar to yours: the AI revolution is real.

Despite abandoning my professional investing career to go long and writing about it for the past three years, I still feel I haven't gone long enough. Like you, I look around and see people becoming incredibly wealthy simply by stringing together tokens for AI applications (or investing all their money in the infrastructure projects that generate those tokens), and it sends chills down my spine and fills me with envy. This then creates a feedback loop where I can't tell if my perspective is influenced by envy or if envy is telling me something I already know: "Keep going long."

To some extent, I do feel that "the future is here, and we need massive computing power," so you would definitely want to buy these assets.

I don't think software stocks are performing that well, and the market is selling them off , so there's not much profit to be made there.

Like you, I've also noticed the extremely low valuations of South Korean stocks and am very interested in the opening up of their market, which is clearly related to the recent stock market rally.

I was also surprised by the official quiet relaxation of the supplementary leverage ratio (eSLR), allowing banks and funds to hold less regulatory capital to buy US Treasuries. This is a classic example of massive monetary easing in sheep's clothing.

I can imagine a day when interest rates rise enough to shut down this "liquidity feast," but that time is not yet here.
I can also imagine that war will end this feast; the violent fluctuations there shook me out of the upward trend, so who knows what the future holds?
I can also imagine that Canadian bank stocks with a price-to-book ratio as high as 3 and very low volatility would be an excellent short-selling opportunity, but due to the lack of trading channels and sufficiently long-term options, I can't write a good article to provide you with some practical information.

Frankly, there are many things I can't say explicitly here. While this won't change my fundamental view on trends, it certainly greatly limits the people and things I can discuss here. If you're familiar with Andreesen's "stop internal friction" theory, you'll know that my overly cautious nature means I'll never become a billionaire.

However, there's one thing I know how to do. This is also a small alpha benefit I can offer you. Today, we're not discussing whether we're in a bubble, but rather how you can short a bubble if you want to.

Why is it so difficult to short a bubble?

What is a bubble? If something looks like a bubble, sounds like a bubble, its price trajectory is like a parabola soaring into the sky, and it requires increasingly higher expectations and leverage to sustain its price increase, then it is a bubble.

Why are bubbles so difficult to short?

The problem is that the easiest things to short are those whose fundamentally negative factors gradually become known to the public, leading to a steady decline and eventual collapse. During this process, you might encounter a short squeeze (short sellers are forced to buy to cover their positions, causing a surge), but this actually provides you with a good opportunity to add to your short position, because this thing will eventually go to zero.

But shorting a bubble is a completely different matter. When an asset price rises unsustainably, your short exposure will increase exponentially as the price rises.

If you don't believe me, go ask the people who shorted Porsche and Volkswagen in 2008.

Ask the people who shorted GameStop.

Or ask the people who shorted that unknown shoe company a few weeks ago that inexplicably turned into an AI company and crushed all the short sellers.

If someone who has gone long sells, they can simply remain out of the market and wait and see. But if someone who has gone short sells, it means they must buy back the next day to close their position and settle their account. If you can make their account five times larger, they will have double the incentive to close their position, sometimes even at any cost.

Another reason why bubbles are difficult to short is that it is precisely those characteristics that make bubbles look so appealing—"skyrocketing volatility! Fantastic!"—that make their options ridiculously expensive.

If it rises by 10% every day, the annualized volatility is 160. For options with a volatility as high as 160, just buying a call option today would cost you the equivalent of half the stock price. Because the hedging value brought by actual volatility is too high, these options are simply unusable for betting on one direction only.

Therefore, we are left with only the following paths.

The only way to short a bubble is:

a) Find the “wedge” – find something that can puncture the foam from the outside.

b) Shorting “victims” – betting on things that are associated with bubbles and whose fall is bottomless.

c) Wait for "confirmation" - wait for the trend and chart to truly break out.

The remainder of this article will provide examples of each method.

A) Finding the wedge

The first way to short a bubble is not to short the bubble itself.

You need to find something that can burst the bubble. Then you buy it to protect your account from the shock of the bursting bubble.

We're starting to do it today, even before the CPI (Consumer Price Index) data confirms what we already knew: inflation is on the rise.

Interest rates are likely to rise as well. As Bob Prince has often said, stocks do indeed possess bond-like characteristics.

This is the "wedge." You shouldn't short the bubble; you should go long on the trend that will kill it. If AI is a bubble, then interest rates are the wedge that punctures it.

All assets with outrageously high valuations are essentially disguised ultra-long-term assets. When the discount rate (interest rate) rises, the discounted value of those promising future expectations will be greatly reduced, and those stocks that have been inflated based on fantasies about cash flows in 2030 will be exposed for what they truly are.

The core principle is this: in every bubble, there are things that depend on the bubble to survive. If the bubble pauses even slightly, the weakest link will break. You're not betting that the market frenzy will end; you're betting that the weakest link won't survive the market pause.

The beauty of the "wedge" strategy is that you don't need precise market timing. The bubble doesn't even need to burst; it just needs to stop accelerating for a quarter, and those highly leveraged junk assets will start to collapse.

Where is the "wedge" now? Let me tell you what I'm watching. Those Canadian banks with price-to-book ratios as high as 3 are holding a lot of "negative amortization" mortgages (meaning that the money borrowers pay is not even enough to cover the interest, and the difference is directly rolled into the principal, just like PIK loans with interest capitalization). The real estate market they are facing makes the US housing market in 2007 seem incredibly restrained.

I can't buy the options for the banks I want, but I've been keeping an eye on them. Speaking of the broader credit market, as we wrote in "Observing Credit," the current private lending market feels like a "cockroach trap," reflecting increasingly lax overall lending standards. Once the money is invested, it can't get out. When the bubble bursts, the book value of these assets won't change because no one is forcing them to revalue. Until the day they have to face reality.

B) Victims of short selling

The second way to short a bubble is to find what will be taken down with it when the bubble bursts, that is, the assets right next to the bubble.

Evergrande is a prime example. You don't need to short Chinese bank stocks; that will only cost you ten years of money. What you need to look for is a developer with ridiculously high leverage, heavily reliant on pre-sales of properties, capable of causing even a slight slowdown in the Chinese housing market to cause it to explode. Bubbles can continue to inflate, but Evergrande couldn't withstand that.

What you need to look for is "downward convexity" (that is, assets that fall faster and faster, and at increasingly larger magnitudes). You can't directly short those that are experiencing exponential growth; that would be fighting against double the upward momentum.

But look at its neighbors; perhaps its options volatility isn't as exaggerated as it's been driven up to 70.

Think back to airlines before the pandemic. They weren't inherently in a bubble, but because they faced extremely asymmetric risks, their crashes could be devastating. Put option prices were high, but not outrageously so. You could still buy options on both ends of the spectrum. So that's what we did. In hindsight, this seems obvious, but the "bubble" at the time was actually just blind optimism about "everything being normal."

Let's recall the financial stocks of 2007/08. You didn't need to directly short real estate (frankly, directly shorting real estate is extremely difficult and has a very high technical barrier; of course, if you could actually find a mortgage-backed securities (CDS) default swap, that would be impressive). You only needed to short Bank of America.

The core principle is that bubbles create a correlation that only becomes apparent during a crash. The options market typically only prices this correlation when disaster strikes. Your task is to identify those "victims"—those with cheap options who are destined to be dragged down by those bubble assets with expensive options.

As for who the current "victims" are? To be honest, I haven't figured it out yet.

C) Awaiting confirmation

The third method requires the most discipline, which is why most people mess it up.

That is: wait.

I know that waiting is the most agonizing part. Sometimes when you see something skyrocketing in price, you just can't control yourself. But let me repeat, you definitely don't want to be run over by a train moving at full speed.

So you need to wait for a confirmation signal. What does the signal look like?

It is usually a combination of the following:

The fundamentals are starting to deteriorate;
Buying activity has dried up, and market sentiment has been exhausted.
The trend line has been completely broken.

Note that this is not a minor pullback, but a complete breakdown. It's the kind of breakdown where something was doing well, suddenly breaks through a strong support line, and everyone starts frantically screenshotting and sharing it on Twitter. We saw this kind of breakdown in silver's price movement in January of this year (but don't look at it now, it's rebounded; we'll discuss this in a future article).

The information presented by the chart will vary depending on the time period you are viewing.

The core truth right now is that the only thing that's getting worse regarding AI is that too much of its cash flow is tied up in the distant future.

The problem is that you have to discount the future pie with today's interest rates. If inflation picks up and policymakers are forced to tighten monetary policy (imagine if oil prices soar to $150 to $200 a barrel, they certainly would), then the net present value (NPV) of many of these assets will shrink dramatically. This is exactly the same logic we wrote about during the bond bubble in 2021.

Another factor to watch out for is relevance. Be wary when previously foolproof methods suddenly stop working, or when they suddenly become sensitive to factors that were previously easily ignored. We may be witnessing this situation today.

Practice and Summary

What did I do today? (Early morning of May 13th, Beijing time) Before the market crash, I had done some hedging, but it wasn't enough. I shorted the S&P 500 (SPX) by 5% and high-yield bonds (HYG) by 10%, and then bought some short-term put option spreads. Then I left for a while, and when I came back, things were bad.

What exactly did I do? I didn't short semiconductors because the core fundamental demand is still there, and the upward trend hasn't been broken. But I did short more bonds, this time directly buying put option spreads on US Treasury bonds. If the trend line holds and the market rebounds, I'll consider it a small expense to reassure my "wedge" strategy—no big deal. If the trend line doesn't hold, I still have cash on hand and protective positions, and only then will I heavily invest in those specific shorting targets. Oh, and I also sold 5% of my Canadian bank stocks.

Hedging, finding a wedge, waiting for confirmation, and then making a heavy investment.

Listen, I don't know if we're in a bubble right now. This rally might only be in its fourth phase (most likely not, the price action has been too strong), or it might already be in its ninth phase (I don't really believe that; that would require the market's demand for tokens—the underlying computing power—to be destroyed, but I haven't seen any signs of that yet). The only thing I do know is that the "unstoppable" feeling that AI gives me is very similar to how I felt when I first cobbled together my internet stock portfolio in high school in 1999. Yes, those stocks eventually recovered, and giants like Amazon were born from them; if you had held them until today, your internal rate of return (IRR) would be in the teens.

But I haven't forgotten that devastating crash back then either.

So, if you've read this rambling essay to the end, you might be feeling nervous. If you are, the answer is definitely not to short the thing that's rising vertically. The answer is: find the wedge, buy the victim's put option, wait for confirmation, and then go all in.

During this period, never go against the market trend. Never short anything that is experiencing a parabolic surge.

ByteDance plans to increase its AI infrastructure spending by 25% to 200 billion yuan.

Popular Articles