Half-year Trading Review and In-depth Market Thoughts Part 2 ～Analysis on Whether the AI Industry Has a Bubble

The current AI bubble is by no means a repeat of the 2000 internet bubble $Invesco QQQ Trust(QQQ.US) $Micron Tech(MU.US) $Meta Platforms(META.US)

As the world's leading AI companies release their earnings reports in a concentrated manner, the industry is significantly ramping up capital expenditures. Coupled with the recent deep correction in the Korean memory sector, market debates about "AI repeating the 2000 internet bubble" have heated up again.

It is undeniable that there are indeed bubbles in the current AI sector, but they are phased, structural, and localized bubbles, fundamentally different from the purely speculative, systemic crash of the 2000 internet bubble. The two cannot be mentioned in the same breath.

1. Reviewing the 2000 Internet Bubble: Rootless Pure Speculation

The core issue of the 2000 internet bubble was the complete lack of a logic for commercial implementation. At that time, the vast majority of internet companies lacked mature profit models and had no real revenue or profit support. They could obtain sky-high valuations with price-to-sales ratios of hundreds based solely on domain names and business PowerPoint presentations.

The entire industry fell into a vicious cycle of blindly burning money for expansion, with all funds consumed on marketing and scaling up, without any sustainable monetization path. After the bubble burst, over 90% of internet companies were completely eliminated and vanished, leading to a systemic collapse of the entire sector.

2. The Current AI Rally: Grounded in Technology Implementation, Possessing Real Industrial Value

The underlying foundation of the current AI industry is extremely solid, having completely shed pure speculation. It possesses tangible technological value and commercial capabilities:

1. Mature Commercial Profit System
The entire AI industry chain has formed a clear monetization model. Computing power chip companies, with NVIDIA at the core, achieve hundreds of billions in annual profits through GPU supply. Tech giants like Microsoft, Google, and Amazon have successfully commercialized their generative AI tools and enterprise-level AI services, continuously generating stable revenue and profits.

2. Tangible Value of Productivity Transformation
AI is no longer a castle in the air; it has deeply penetrated industrial scenarios. Applications such as automated code generation, intelligent customer service, data analysis, and enterprise cost reduction and efficiency improvement have been fully implemented, genuinely enhancing social and corporate production efficiency, possessing irreplaceable industrial value.

3. Ultra-High-Speed Market Penetration
Compared to the lengthy adoption cycle of the early internet, the current acceptance speed and implementation penetration rate of AI by enterprises and industries far exceed those of the past. The pace of industrial adoption has accelerated comprehensively, and market acceptance has been genuinely verified.

4. Robust Capital Investment Structure
The 2000 internet industry relied on external investment funds for money-burning expansion, with extremely poor financial risk resistance. In contrast, the AI infrastructure investments by leading companies today mostly come from net profits accumulated through their own operations, resulting in a more stable financial foundation and stronger sustainability.

3. The Core Issue of the AI Structural Bubble: Overdrawn Expectations, Supply-Demand Mismatch

Although the AI industry foundation is solid, the problems of industry overheating and localized bubbles are very prominent. The core contradiction lies in the severe mismatch between aggressive investment pace and lagging commercial returns.

Currently, global tech giants' annual AI capital expenditures reach $700-900 billion, but the new direct revenue generated by AI is only in the tens of billions, creating a huge gap between input and output. Data shows that over 90% of generative AI pilot projects have so far failed to generate quantifiable profit contributions.

This round of bubble exhibits strong structural differentiation characteristics: upstream computing power chips and infrastructure service providers continue to reap huge profits with fully realized performance, showing almost no bubble; midstream and downstream AI application companies are generally stuck in continuous money-burning difficulties, with severely inflated valuations.

The real risk for the industry is not that AI technology lacks value, but that the market has excessively overdrawn short-term return expectations: the pace and scale of capital investment are too fast and too large, while the speed of commercialization and profit realization falls far short of capital expectations.

4. Conclusion: AI Will Only Reshuffle, Not Crash

The 2000 internet bubble was a holistic bubble characterized by concept-first, no business model, and no industrial foundation. After it burst, the entire sector collapsed, and companies died in batches.

The current AI industry is experiencing structural overheating, with real technology implementation and established industrial value, but also capital overheating, overdrawn expectations, and lagging returns.

In the long run, AI is a core hard technology that will disrupt the era, and its long-term growth logic is unquestionable. In the short term, the industry is highly likely to undergo a deep correction, washing out inferior players without core technology or profitability through valuation adjustments, completing an industry reshuffle. However, relying on solid technology implementation and industrial demand, the AI sector will definitely not repeat the systemic collapse of the internet era.