Is The AI Bubble Going To Burst?

There are some good points to be made that, in its current form, the AI mania is unsustainable.

TL;DR

AI is here to stay and revolutionize the world, there’s no arguing that.

But that doesn’t mean AI is not a bubble.

In today’s post, we discuss both the arguments in favor and against AI being a bubble.

Most importantly, we look at possible warning signs that the bubble could be deflating so that we may exit in time.

Expect To Learn:

• How Bubbles Are Formed

• Differences Between AI and the Internet

• Similarities Between AI and the Internet

• Warning Signs That A Bubble Is Deflating

But first, a little bit about me, The Pragmatic Investor

---

An approach that assesses the truth of meaning of theories or beliefs in terms of the success of their practical application.

That is Pragmatism, and it guides my investment philosophy.

Through many years of analyzing markets, I have found this is what works.

The Pragmatic Investor sticks to what works. The problem is there are thousands of different strategies that CAN work for certain people.

But simplicity is key, and that is why I have narrowed down my investing ethos into three key ideas, which combined create what I like to call, The Pragmatic Investing Pyramid.

Macro, Fundamentals and Technical.

This three-pronged approach has helped me beat markets over the seven years.

For long-term investing, there’s nothing better than understanding business cycles, macroeconomic trends and geopolitics.

On the other hand, when it comes to short-term moves in markets, the best tool we have is technical analysis.

And not just a specific form of technical analysis but a robust set of tools that can all work in conjunction to help us find great setups.

I actually recently designed my own algorithm, you can see it here:

My Substack is designed to guide investors of all levels in their journey.

• Understand markets with the weekly newsletter.

• Build a diversified global portfolio that will stand the test of time.

• Get actionable trade ideas to take your investment returns to the next level.

Every week I give a macro update, a technical analysis update on the main indexes and stocks and I cover a stock in-depth, looking at its fundamentals.

The AI Bubble Is Unsustainable

There are some good points to be made that, in its current form, the AI mania is unsustainable. While no one can deny that AI technology is groundbreaking, the same could be said of the Internet, and yet this also created a massive bubble.

When all was said and done, the Nasdaq lost 83% of its value. If that were to happen today, we’d see the Nasdaq back to almost 3500.

Since the internet is perhaps the most comparable technological revolution to AI in recent history, it could serve us to compare these two technologies in terms of adoption, investment and profitability in order to answer the question; is AI also a bubble?

Internet vs AI Users

It’s been argued that the adoption of Generative AI is on track to match and even outperform that of the internet.

But is this a fair comparison?

Consider that becoming an “AI user” today comes with close to 0 cost. Meanwhile, becoming an internet user would have required purchasing a computer and access to the internet. Both are costly investments.

In this regard, saying that the use of AI has become widespread is not enough to justify all the hype.

Productivity vs Cost

What we really must look at is the productivity and the cost of AI. What are the benefits? And how much investment does it require?

AI, unlike the internet, is proving to be far more expensive and slower to deliver measurable returns. Deploying AI models demands highly specialised hardware and enormous energy resources, pushing operational costs beyond current capacities. In contrast, the internet was comparatively inexpensive to implement and delivered rapid, tangible benefits.

By 2000, over 134 million PCs had been sold, generating $684 billion in revenue at an average price of $5,101 each. E-commerce alone accounted for $286 billion annually, and every internet user was paying directly for access. Adjusted for inflation, the internet was generating around $1.5 trillion by the year 2000, just before the dot-com bubble burst.

In comparison, Generative AI is projected to generate $10 billion by 2026. However, as David Cahn points out, AI would need to achieve $600 billion in annual revenue to justify its current investment levels—a target that remains far out of reach.

Unsustainable Valuations

The issue here, like with the internet is timing. Everyone can see the potential of this technology and is rushing to invest in it now.

This is leading to valuations which may be unsustainable, especially if AI doesn’t deliver soon.

NVIDIA’s valuation is underpinned by substantial and growing revenues, but the same cannot be said for many AI startups.

For instance, OpenAI, valued at $157 billion in its latest funding round, is projected to incur $5 billion in losses against $3.7 billion in revenue.

The real issue lies not with NVIDIA itself but with the fragile ecosystem it depends on. AI startups are its primary chip buyers, and these companies often rely heavily on venture capital and private equity funding to stay afloat. If one major player collapses, it could trigger a domino effect, drying up funding and potentially causing the entire ecosystem to unravel.

Subscribe to The Signals Doctor:

Is This Time Different?

But we could also argue that there are key differences between AI and the internet bubble. They are summarized in the table below.

The rise of AI today is indeed concentrated in large tech companies. Yes, we are seeing massive investments by Google and Microsoft, but these companies have money to spare. Even if AI became a nothing burger, you would still be left with profitable companies with strong balance sheets. Meanwhile, the internet bubble gave us insane valuations in names like Pets.com.

On top of that, AI is a technology which could transcend into numerous fields across the economy.

Analysts project a compound annual growth rate (CAGR) of 35.7% from 2024 to 2030, reflecting vast untapped potential. If we analogise this bull market to a baseball game, we’re still in the third or fourth inning, with significant room for growth ahead.

How And When Will The AI Bubble Burst?

So what’s going to happen with AI?

In short, I do believe we will witness a bubble, but the fallout won’t be as bad as the .com crash.

Firstly, let’s understand the anatomy of a bubble.

What 's A Bubble?

1. Displacement; Stealth Phase

The bubble begins with a groundbreaking innovation that captivates investors. For AI, this displacement stems from its transformative potential, epitomised by tools like ChatGPT, which showcase the technology's ability to revolutionise industries.

2. Boom; Awareness Phase

As excitement builds, investment pours in. AI has witnessed an unprecedented surge in funding from individuals and corporations eager to capitalise on its promise. Startups and established players alike have scrambled to carve out a share of the AI gold rush.

3. Euphoria; Mania Phase

In this phase, valuations often outpace reality. Companies involved in AI are experiencing “irrational exuberance,” with some achieving valuations that vastly exceed their current revenue-generating capacity. Speculation and hype dominate, and risk becomes secondary to potential rewards.

4. Profit-Taking: Blow-off Phase Pt 1

Astute investors begin to offload assets, recognising the overvaluation. These early movers aim to lock in profits before the market turns. This stage can be subtle and may not immediately signal broader panic.

5. Panic and Collapse; Blow-off Phase Pt 2

When the market recognises the discrepancy between inflated prices and actual value, a rush to sell ensues. Prices plummet, leaving late-stage investors with substantial losses. For AI, this could occur if growth slows or financial returns fail to meet expectations.

But the real question is. How can we be the astute investors that ride the wave but exit in time?

The key is to look for the signs that the bubble could be coming to an end.

Predicting the AI Bubble Burst: Key Warning Signs

Timing the burst of a market bubble is notoriously difficult, particularly in fast-moving sectors like artificial intelligence. However, several indicators can signal when AI stocks might lose momentum, paving the way for a potential market correction:

1. Disappointing Financial Results

When major AI companies fail to meet revenue or profitability expectations, it erodes investor confidence. Missed earnings targets, slower-than-expected growth, or widening losses could lead to sharp declines in stock prices, creating a ripple effect across the AI sector.

2. Waning Investor Interest

High valuations in bubbles are often driven by speculative enthusiasm. A shift in sentiment—whether due to fatigue with AI or the emergence of another "next big thing" in tech—could result in a significant pullback. Diversion of capital to safer investments or competing technologies could further deflate the AI bubble.

3. Technological Limitations and Overhyped Expectations

The gap between AI’s potential and its current limitations is another potential pressure point. If advancements stall, scalability proves elusive, or highly publicized projects fail to deliver, investors may become disillusioned. This disillusionment could trigger a widespread sell-off, exposing overvalued companies.

4. Increased Government Regulation

As governments seek to address privacy, security, and ethical concerns in AI, increased regulation could disrupt the industry. Stringent compliance requirements, restrictions on data usage, or limits on AI applications could stifle innovation and profitability, discouraging investors and slowing growth.

5. A Broader Economic Downturn

AI is a high-risk, high-reward sector, making it particularly vulnerable during recessions or economic slowdowns. In such scenarios, investors often shift to more stable assets, causing sectors like AI to experience sharp corrections as capital dries up.

6. Major Industry Scandals or Failures

Unexpected events, such as high-profile legal challenges, data breaches, or technological failures, can shake confidence in AI as a whole. A single prominent company’s collapse or scandal could act as a catalyst, triggering broader market panic.

Final Thoughts

We’ve definitely seen some, if not all of these signs occur.

While NVIDIA has not yet disappointed investors, its latest earnings were underwhelming, and growth is slowing.

We are also seeing increased scrutiny by governments, rightfully so, and this could pose a threat to the industry.

And, we have even had major scandals, such as SMCI’s dodgy accounting practices, which might even get the stock delisted from the NASDAQ.

Under the right macroeconomic conditions, AI could become too costly to keep investing in, and we could see this whole thing unwind.

If AI goes, this will drag the market down with it. Never before has the market been so concentrated in just a few companies.

For this reason, tracking macroeconomic conditions is perhaps my top priority as a Pragmatic Investor.

Thanks for reading, and be sure to check out my substack if you liked this article.