Is AI a Bubble? Bill Gurley Uncovers the 'True Nature of Industrial Bubbles' and the Reality of China, Careers, and Investment

The frenzy surrounding AI is expanding at an unprecedented speed, involving investors, entrepreneurs, and policymakers alike. Some dismiss this situation as a 'bubble,' while others assert it is the 'beginning of an industrial revolution.'
However, this dichotomy itself is already close to a cessation of thought.

Bill Gurley, who has experienced multiple technology waves and market crashes since the dawn of the internet, views AI, China, and individual career choices not as mere trends, but as issues of 'structure.' Based on his interviews, this article organizes how we should understand the current AI boom and where we should stand.

1. Is AI a bubble? — 'Technology waves' and 'speculation' always appear in pairs

1-1. The question 'Is AI a bubble?' is wrong in itself

Is AI a bubble, or is it a genuine technological revolution?
Bill Gurley says that the way this question is framed is fundamentally flawed.

He relies on Carlota Perez's theory. Looking back at history, large-scale technological innovations such as the steam engine, railroads, electricity, automobiles, and the internet have all proceeded simultaneously with excessive financial speculation without exception. In other words, speculation concentrates precisely because it is technically genuine, and the existence of a bubble is not evidence against the technology.

If a gold mine is discovered, serious miners and those looking to strike it rich gather at the same time. AI is the same. Just because speculation exists does not mean the essence of the technology is denied. In fact, a technology that generates no speculation at all is highly unlikely to have the power to change society.

1-2. The danger of 'believing/not believing' binary thinking

What Gurley is most wary of is the simplistic argument that 'people who say AI is a bubble are denying AI itself.' This way of thinking prevents us from grasping reality correctly.

What is important is to recognize the dual structure: that AI is a genuine technology that will change social structures, and at the same time, excessive expectations and speculation are inflating in the capital markets. Technological progress and market overheating can exist simultaneously.
If you cannot accept both at the same time, you cannot make a calm judgment.

2. AI as an 'industrial bubble' — The decisive difference from the dot-com era

2-1. The 'true legacy' left by the dot-com crash

Gurley views the AI boom as overlapping with the internet bubble of 1999 to 2000. However, what is important here is to discard the misconception that 'bubble = meaningless.'

Jeff Bezos has said that there are two types of bubbles. One is a financial bubble, and the other is an industrial bubble. The 2008 housing bubble was the former, and it left almost nothing for society after it collapsed. On the other hand, the internet bubble was the latter. Many companies disappeared, but the foundations—fiber optics, data centers, protocols, and entrepreneurial spirit—remained, and they gave birth to the GAFA of today.

AI clearly belongs to this 'industrial bubble.' Excessive investment will be corrected, but computing resources, models, data, and business designs based on AI will not disappear.

2-2. Yet, the distortion of 'circular trading' cannot be overlooked

On the other hand, Gurley also points out the dangers unique to the current AI industry. The symbol of this is 'circular trading.'

Major tech companies make huge investments in AI startups, and that money returns as usage fees for their own cloud services. This structure looks rational at first glance, but in reality, it has the effect of artificially inflating sales and valuations.

Gurley's question is simple.
If the transaction is not important, why do it?
If you admit it is important, how can you assert that it does not undermine the health of the market?

If this structure becomes normalized, the boundary between investment decisions and actual demand will become blurred, and the final adjustment will arrive in a more painful form.

3. The Trap of Individual Investors and SPVs: 'The 100x Days Are Over'

3-1. Why SPVs Are Dangerous

Gurley specifically warns about the trend of individual investors participating in AI investments through SPVs. The structure of SPVs is complex, and in many cases, investors do not actually hold the shares directly.

It is more like a 'right' to acquire shares in the future, fees are high, and information disclosure is limited. Furthermore, there is little freedom to sell. Gurley describes this structure as the 'Wild West.' Rules are underdeveloped, and investors in weaker positions are the most disadvantaged.

3-2. Finding 'the Next OpenAI' Is Nearly Impossible Now

Another reality is the expected return. 100x or 1000x returns only occur in the early stages when a company is unknown. AI companies that everyone knows by name today have already passed that phase.

Furthermore, in the world of VC investment, not just a majority, but 60% to 80% end in failure. Many people overestimate their own risk tolerance. It is only when their assets are actually halved that people truly learn their limits.

4. The 'Misunderstood Competitive Society' Seen While Walking Through China

4-1. China Is Not a 'Non-Competitive Society'

China has a planned economy and no competition. Gurley says this perception is held most strongly by those who have not seen the country firsthand.

While it is true that the central government sets five-year plans, their execution is left to local governments. And local regions compete fiercely over attracting investment, fostering industries, and growth rates. If they produce results, they are promoted; if they fail, they are weeded out. This structure is, in fact, more competitive than that of the United States.

4-2. The Xiaomi SU7 as a 'Mass of Execution'

What shocked Gurley the most was Xiaomi's EV project. Founder Lei Jun has 200 of his own employees drive prototype cars and requires them to write down areas for improvement. He immediately reflects that feedback in design and manufacturing.

The EV factory is highly automated, and the number of employees is overwhelmingly lower compared to a US factory of the same scale. What we have here is a reality that shatters the illusion that 'bringing manufacturing back home will increase employment.'

5. The Core of National Competitiveness: What Is America Losing?

5-1. Is Market Capitalization Really National Power?

The Chinese government does not prioritize corporate market capitalization as an indicator of national success. What is prioritized is price competitiveness, export capacity, and industrial infrastructure.

Meanwhile, in the United States, the market capitalization of giant corporations is often spoken of as a symbol of national power. However, Gurley raises a question: How much benefit does the existence of a $3 trillion company bring to society as a whole?

5-2. A 'Nation of Lawyers' vs. a 'Nation of Engineers'

Gurley symbolically describes the difference between the two countries as follows: America is a nation of lawyers, and China is a nation of engineers.

Nuclear power plants, factories, and semiconductor manufacturing. In the United States, regulations and lawsuits significantly slow down the speed of construction. If we are serious about re-industrialization, we must be prepared to return to being a 'country that can build.'

6. 'Running Down a Dream'—Why do people give up on their dreams?

6-1. Permission not to do things halfway

The words actor Matthew McConaughey was told by his father, 'If you're going to do it, don't do it halfway,' are symbolic for Gurley as well. What these words provide is not an obligation to work hard, but permission to give it your all.

Whether or not you can grant yourself this permission becomes a turning point in life.

6-2. Passion can be measured by 'how you spend your free time'

Gurley presents a simple standard for measuring passion: whether or not you are learning on your own time without being told to by anyone.

Successful people, without exception, continue to learn during time they are not forced to. It is not a matter of talent, but a matter of habit.

7. Career Strategy in the AI Era—How to stand on the 'side that won't be replaced'

7-1. The humans closest to AI will survive

Gurley asserts: the best way not to have your job taken by AI is to stand on the side that uses AI the most effectively. Single skills are dangerous, but roles redesigned with AI as a premise actually increase in value.

7-2. Regret is born from 'things not done'

As Jeff Bezos's regret-minimization framework shows, people regret what they did not do the most. This principle is universal, transcending culture and country.

Conclusion: Pursue dreams through 'design,' not 'momentum'

The worldview Bill Gurley speaks of is based on structural understanding, not emotional arguments. Technology, nations, and careers are all things that should be approached through long-term design rather than short-term frenzy.

The resolve not to do things halfway and the attitude of continuing to learn. Perhaps those are the only conditions for continuing to run through an uncertain era.

Recommended Articles

Next Big Wave (Growth Stocks, Seeds of Ideas, Deep Dives into Trends)