The real dot-com bubble is repeating itself with Web3, not AI.

Stock market and cryptocurrency investors are worried that AI will repeat the dot-com bubble of 2000. In fact, Web3 is the real dot-com bubble repeat.

Bubble: Supply-side self-indulgence, weak demand

During the dot-com bubble of 2000, funds were indulging in a self-indulgent frenzy on the supply side. Many listed internet companies had no profits or cash flow, yet their stock prices soared.

On the other hand, there are very few internet applications on the demand side... 2,000 years ago there was no Douyin, Kuaishou, Alibaba, JD.com, or WeChat.

At that time, the mainstream internet applications were only chat software such as OICQ (the prototype of QQ) and MSN, download tools such as FlashGet and NetAnts, information platforms such as Yahoo, NetEase, and Sohu, basic applications such as Google Search and email, and Amazon and eBay, as early stages of e-commerce, had very few users.

Back then, there was no mobile internet, and personal computers were a luxury. Most families in China didn't own a computer, and the monitors were those bulky CRT displays. Laptops were as thick as bricks, and internet access required dialing via a telephone line. Even older were the storage media; the primary storage medium was the floppy disk—yes, the A drive. Its capacity was only 1.44MB, not even enough to hold a slightly larger image today.

It's worth mentioning that there were single-player games at the time, but they were pixelated games.

AI: Infrastructure leaders surge, products in growth phase

Looking at the AI ​​industry, while stock prices have surged, Nvidia, AMD, SK Hynix, Samsung, Micron… these are primarily AI leaders, and in the infrastructure sector. This is because AI requires extensive training before it can be deployed. On the supply side, the AI ​​market doesn't have the exaggerated bubble seen in 2000. At least these leading AI companies do indeed have considerable profits.

On the demand side, we can see at least Gemini, Claude, GPT, Doubao, and various other UGC applications. There's also the recently viral lobster, the newly launched Perplexity Personal Computer, and the eye-catching AI robot on the Spring Festival Gala. Although the breadth of internet products cannot yet be compared to that of 2000, this is because the development of AI requires high-level infrastructure. AI needs to develop its infrastructure first before it can be integrated into more types of applications.

Web3: Technology as a gimmick, with very few applications.

╰┈✦ Market capitalization and earnings

Looking at the Web3 industry, various technical narratives emerge one after another, but how many applications actually have users? Apart from a handful of DeFi leaders, there may only be MEME platforms, prediction markets, and Perp dex, and much of the activity in the latter two still comes from interactions based on anticipated airdrops.

The shortage of products on the demand side corresponds to the self-congratulatory attitude on the supply side.

⏵ An example from the ZK sector: ZKsyn, with a daily revenue of $458, has a market capitalization of $176 million, equivalent to a PE ratio of 1052.

⏵ An example from the L2 sector: Optimism, with a daily revenue of $2,427 and a market capitalization of $253 million, has a PE ratio of 285.

⏵ An example from the L1 sector: Sei, with a daily revenue of $3,564, has a market capitalization of $424 million, equivalent to a PE ratio of 327.

There's also Blast, which offers a daily income of -$6...

If we follow the logic of the stock market, shareholders of the first three projects would have to wait 1052, 285, and 327 years respectively to break even, not even taking into account the infrastructure and operating costs required to maintain these chains.

Although these ecosystems don't rely on on-chain revenue to recoup costs and generate profits, for shareholders, i.e., token holders, it's a nightmare...

What are some Web3 applications?

Let's take a look at some Web3 applications.

Metaverse, blockchain games, inscriptions, social networking... and perhaps some other applications we can't even remember, are now almost never mentioned.

Currently, besides DeFi and RWA, the only other platforms with users are memes, prediction markets, and Perp DEX. Among these, memes involve existing funds in PvP, while prediction markets and Perp DEX are partly driven by interactions based on airdrop expectations.

Compared to the internet applications mentioned earlier in 2000, Web3 applications were truly few and far between…

In conclusion

Therefore, what is truly indulging in wishful thinking on the supply side while lacking any real progress on the demand side is not AI, but Web3.

The players in the US stock market overlap with those from 26 years ago, and investors and Wall Street are deliberately avoiding a repeat of history. There is a bubble in AI, but it differs from the dot-com bubble of 2000.

On the contrary, in the younger Web3 field, capital is speculating on technology on the supply side, while there are not many truly practical products on the demand side. This is a repeat of the dot-com bubble of 2000.

Based on AI-driven moderate bubble and Web3 high bubble inference:

First, the US stock market will correct, but the possibility of a crash like in 2000 is unlikely.

Second, BTC, which is linked to US stocks, was hit to a moderate degree.

Third, just like the dot-com bubble of 2000, altcoins will continue to undergo a shakeout, forming a painful process of weeding out the fakes and retaining the genuine ones. This process may take longer than I imagined.

The overall decline in counterfeit goods has lasted for about 15 months since the end of 2024, but this is not the end.

Don't believe the claim that counterfeiting has reached its limit. Some counterfeiting has reached its limit, but others have no limit.