The truth about the AI bubble revealed: 244 times the price-to-sales ratio, far exceeding the internet peak. Is it an opportunity or a trap?

As the AI boom sweeps across the globe, we set aside emotions and focus on the data itself. Taking ChatGPT as an example, among 800 million monthly active users worldwide, only 37.5 million are paying users, generating approximately $11.25 billion annually. Extending this to the entire AI consumer market, assuming each person subscribes to 3 services, the total scale reaches $33.75 billion. In comparison, the combined market capitalization of the seven AI giants is $19.8 trillion. Even if only 25% of this is attributed to AI premium, it amounts to $8.25 trillion, with a price-to-sales ratio as high as 244—far exceeding the valuations of top internet companies during the 2000 dot-com bubble. Does this indicate a bubble burst? Let’s calmly do the math and see the true value and potential risks of AI.

Calm Breakdown of the AI Market Size

ChatGPT and Consumer Market Revenue Estimation

Based on publicly available data, we start with ChatGPT, estimating its consumer revenue and extrapolating to the entire AI market. The following table shows the key calculations:

This estimate is quite optimistic; many users subscribe to only 1-2 services. It’s difficult for ordinary consumers to maintain more than 3 subscriptions long-term, and even heavy users like myself spend around $100 per month.

AI Seven Giants’ Market Cap and Valuation Premium

In the secondary market, the narrative around AI directly boosts the market caps of the seven major tech giants. We conservatively assume only 25% of their market value reflects AI premium, and calculate the price-to-sales ratio. The table below lists the giants’ market caps (based on the latest data):

The total AI premium of $8.25 trillion divided by the consumer market of $33.75 billion yields a P/S ratio of 244. This is far higher than the peak of the internet bubble in 2000, where Amazon’s P/S was 19 and Microsoft’s 26. Even after a bubble burst, winners like Amazon would need 10-15 years to recover their stock prices.

AI Valuation vs. Historical Bubbles

Apart from the nearly 1 trillion valuation in the primary market, the secondary market already shows signs of overvaluation. During the internet bubble, top companies’ P/S ratios ranged from 16 to 35, whereas AI’s current 244 indicates market sentiment far exceeds reality. While AI is an efficient productivity tool capable of reducing costs for entry-level jobs (companies willing to pay $50-100/month), it is not “productivity itself,” and pricing it at the level of hiring personnel ($2000-3000/month) is challenging. Yet, the capital market assigns it a “revolutionary” valuation, revealing a clear contradiction.

Future Optimistic Scenarios and the AGI Variable

Even in an extremely optimistic scenario, if consumer paid subscriptions increase tenfold to 375 million (exceeding the US population), the P/S ratio would still be 24, which is not cheap for a mature industry. But if AGI is achieved, AI could become a substitute for human productivity, and today’s valuation might become “cheap.” Looking back at Amazon in 2000, from a 2025 perspective, it would seem “ridiculously cheap.” The prerequisite is patience—buy in and wait 10-15 years. When AGI arrives is unknown, so AI is neither an absolute bubble nor permanently overvalued. It’s advisable to maintain patience between the narrative and reality, taking one step at a time.