There is a piece of data worth paying attention to—AI computing power consumption in the US has already accounted for 7% of the country's total electricity generation. This proportion may not seem large, but upon deeper reflection, being just one step away from 8% makes it significant.

What’s even more concerning is that the value generated by this electricity consumption has become a question mark. Let’s see what 8% really means: it’s roughly equivalent to three times 2.714%. In other words, the demand for AI computing power in the US is consuming the natural growth rate of the country’s three major pillar industries.

This is the real issue. Once the power consumption for computing exceeds 8% and still fails to produce genuine value, the entire logic of AI computing power will be challenged. Because this is akin to mobilizing the entire nation’s effort to force the growth dividends of the three major industries into the AI basket, only to find it’s like drawing water with a bamboo basket—value has nowhere to go.

Of course, to continue boosting stock prices, some capital may keep stacking computing power and expanding electricity consumption. But there is a strict rule: 33.33% of the total US electricity capacity is the ceiling. Once this line is crossed, the consequences are no longer just a data game—they will directly evolve into AI encroaching on real life. By then, it’s not just the economic ledger that can’t support it, but the entire functioning of the country.

So the key isn’t how high 7% is right now, but that this number is rapidly approaching a critical point, and we still can’t see clearly what true value AI computing power can bring.