Howard Marks Latest Memo: Is it a bubble? Maybe a ‘useful’ bubble

Howard Marks Latest Memo: Is it a bubble?

The Essence of the Game

In Oaktree Capital’s latest memo, Marks opened by addressing the question every client cares about most: “Is It a Bubble?”

Marks did not give a simple “Yes” or “No.”

Instead, through this memo, he revealed the essence of this capital game: We are at a historic turning point. Market mania might be irrational, but that doesn’t mean it is valueless. Marks believes that to understand the current AI craze, we must distinguish between two distinct types of bubbles. This distinction will determine whether investors ultimately become the fuel for progress or the winners of the future.

The Duality of Bubbles: “Money Burns, but Tech Remains”

Marks cites the theories of Byrne Hobart and Tobias Huber to introduce a deeply insightful concept: “Inflection Bubbles.”

• Mean-reversion Bubbles: This is the traditional “bad bubble,” such as the 2008 subprime mortgage crisis. They are built on financial engineering. When the bubble bursts, aside from wealth destruction, the world has not progressed; everything simply returns to square one.
• Inflection Bubbles: These bubbles are usually accompanied by revolutionary technology (such as railways or the internet). In this scenario, investors’ over-optimism provides massive capital, funding high-risk infrastructure construction (like today’s data centers and chip R&D).

Marks emphasizes that while “Inflection Bubbles” can accelerate technology adoption and lay the foundation for a prosperous future, they also destroy wealth. Even if the technology ultimately changes the world (like radio or aviation in the 1920s), investors who blindly chased highs could still lose over 90% of their assets when the bubble bursts. Simply put: Society progresses, but the investors who paid for it burn through their money.

Danger Signals: Circular Trading & Debt Traps

In the memo, Marks points out two of the most worrying financial signals in the current AI market:

1. Circular Deals:

Marks is skeptical of the cash flows between current tech giants and startups. He notes that companies like OpenAI receive billions in investment from tech giants (like Microsoft, Nvidia), but then use those funds to buy cloud services or chips from those same investors. This “left hand to right hand” trading pattern makes revenue figures look dazzling, but the profits may be illusory, similar to the fiber-optic swap deals during the 90s telecom bubble.

2. The Misuse of Debt:

This is the biggest difference between this bubble and previous ones. Marks warns that AI infrastructure build-outs are relying heavily on debt financing. Many companies are issuing 30-year bonds to fund AI investments where the technology changes extremely fast. He questions the wisdom of using long-term debt to fund assets with highly uncertain outcomes (like chips that might be obsolete in a few years). In a winner-takes-all market, equity investment might cover all losses through one winner, but with debt investment, once failure hits, the losses cannot be recovered.

Comparison with 2000: Fundamentals Are Different

Despite the flashing risk signals, Marks acknowledges that the current AI boom is fundamentally different from the Dot-com bubble of 2000. The most direct evidence comes from valuation data.

Marks presents data from Goldman Sachs comparing the P/E ratios of today’s AI giants with the tech bubble era:

• •In December 1999, Microsoft’s forward P/E was as high as 69x, Cisco was 101x, and Oracle was 90x.
• In contrast, today Microsoft’s P/E is only about 30x, almost half the price of back then.

Unlike the many “.com” companies with no revenue back then, today’s AI products already exist at scale, demand is exploding, and the market is dominated by mature enterprises with strong cash flows. Although market sentiment is manic, the fundamental support is much more solid than 25 years ago.

Maintaining Moderation Amidst Uncertainty

At the end of the article, Marks quotes a golden rule: “Bubbles are best identified in retrospect.” No one can accurately predict when a bubble will burst, nor can they be certain who the ultimate AI winners will be.

Therefore, his advice is not to “flee the market,” but to maintain a “Moderate Position.”

• Do not All-in: Because uncertainty is too high, and there is a risk of permanent capital loss.
• Do not All-out: Because this may indeed be one of the greatest technological shifts in human history, and missing it completely would be a huge regret.

Marks reminds us that when facing this “useful bubble” that might change the world, participating with a selective and prudent attitude is the best strategy for investors dealing with an unknown future.

Battle Tactics: Surviving in the “Fog”

Howard Marks tells us this might be a “good bubble” that changes the world, but don’t forget he also coldly adds a caveat: “Good bubbles also destroy wealth.” History tells us that even though the internet eventually changed human life, many who bought Cisco or Yahoo at the peak back then never made their money back.

We don’t need to be prophets; we just need to be survivors who strictly follow discipline.

1. Quit the “Lottery Mindset,” Strictly Adhere to Stop-Loss Discipline

In the AI frenzy, don’t refuse to cut positions because of “dreams.” If your holding is built on the assumption that “it will rule the universe in the future,” but the stock price falls below key support, please respect the market.

Marks says “debt” is the invisible killer of this bubble, which means once the wind shifts, the price correction for highly leveraged companies will be faster and more brutal than you imagine.

Set strict stop-losses; don’t let one trade take away half your principal.

2. Take Profits Appropriately: No One Goes Broke Locking in Gains

When your AI holdings have doubled, or even make you feel “making money is too easy,” this is the sign of “irrational exuberance” Marks speaks of. At this point, taking profits in batches is the optimal answer.

Don’t fantasize about selling at the absolute peak. Turning paper wealth into cash in your pocket makes you invincible. Remember, the technology will remain, but if you don’t pocket your money, you might just end up being a “philanthropist” for these technological developments.

3. The Middle Way: Core & Satellite Allocation

Don’t bet your life savings on stocks with P/E ratios of 100x. You can use a “Barbell Strategy”: put the majority of funds in quality stocks with substantial cash flow and reasonable P/E ratios (Marks mentions Microsoft is much cheaper now than in 2000), and use a small portion of funds to gamble on high-risk AI startups. That way, if the bubble bursts, you only get a scratch; if the bubble keeps inflating, you can still fly along with it.

I want to strongly suggest to all friends that Howard Marks used an 18-page memo to tell us one thing: Admitting ignorance is the beginning of wisdom.

Since we don’t know when this game will stop, let’s not guess when the music will end. Instead, while the music is still playing, find a spot near the exit, and occasionally put the chips you’ve won into your pocket. This is our way of survival in this crazy era.