Summary of Key Points
Recently, the global AI sector has experienced a significant decline, leading investors to worry about whether technology companies' spending on AI is getting out of control. The market value of major U.S. tech giants lost approximately $800 billion in a single day. Although some companies (such as Alphabet and Tesla) have seen revenue growth, their cash flows are negative. This suggests that the criteria for evaluating AI companies may be shifting from focusing solely on growth to also considering cash flow and capital efficiency. There is disagreement among two economists regarding the impact of AI on the U.S. economy: Freeman believes that AI drives demand and boosts productivity (though not in an explosive manner), while causing minor disruptions in the labor market; Mikhov, on the other hand, argues that AI has not led to a substantial increase in productivity and has had no significant effect on the labor market, suggesting that tech companies' layoffs are merely an excuse for excessive hiring during the pandemic.
Detailed Analysis
1. The Sharp Drop in the AI Sector: Investors Moving from "Concept Chasing" to "Practical Evaluation"
The main reason for the recent plummet in AI stocks is a change in investor sentiment. In the past, investors would buy into AI-related companies at the mention of the term; now, they are more concerned about whether the investments are worthwhile. For example, the significant loss of $800 billion in the market value of the seven major U.S. tech giants indicates that the "free cash flow" (the money remaining after deducting costs) of Alphabet and Tesla has turned negative, meaning their AI investments have not yet generated actual profits. This realization highlights that evaluating AI companies should take into account both revenue growth and the efficiency with which capital is being used.
2. How Does AI Drive Demand in the U.S. Economy?
Freeman believes that AI currently acts as a "demand-side driver." For instance, building data centers requires purchasing land, constructing facilities, and hiring personnel for maintenance, all of which stimulate demand in related industries such as construction, hardware manufacturing, and IT services. As more data centers are established, the demand for raw materials and labor increases, leading to higher prices and employment rates. In other words, AI first "spends money" to stimulate the economy rather than directly increasing productivity.
3. Has AI Really Increased Productivity?
The two economists have opposing views on this:
- Freeman's Optimistic View: He estimates that U.S. labor productivity has grown by an average of 2.1% since 2019, with AI contributing about 0.5 percentage points, and he expects this to rise to 2.5% in the future (though far from the 10% explosive growth claimed by Silicon Valley).
- Mikhov's Cautious View: He argues that the productivity gains expected for 2024-2025 are not due to AI and that in 2026, "total factor productivity" (efficiency improvements stemming from technological advancements) could even be negative, meaning that the same amount of resources may result in less output. He suggests that the Federal Reserve cannot rely on an AI "miracle" to lower interest rates, as such a phenomenon has not yet been observed in reality.
4. The Impact of AI on Jobs: Replacement or Enhancement?
Freeman identifies three types of companies:
- Companies Trying AI but Not Understanding Its Potential: These hire many employees to research AI, but their productivity does not increase, and sometimes even declines due to increased working hours.
- Companies That Have Not Yet Adopted AI: The majority fall into this category.
- A Small Percentage (10%) That Are Using AI Effectively: These see a significant improvement in productivity, though they may experience an initial period of investment without immediate returns, followed by gradual profitability.
Mikhov provides a practical example: Radiologists were once feared to be replaced by AI, but hospitals now have a greater demand for them because AI helps doctors process images more quickly, allowing them to see more patients. Tech companies claim to be laying off employees due to AI, but in reality, these layoffs are often a result of excessive hiring during the pandemic.
5. Future Prospects: A Balance Between Optimism and Caution
- Freeman's Optimism: Even if AI models do not improve further, we are only using a small fraction of their potential, and they will eventually boost productivity.
- Mikhov’s Cautiousness: No clear "miracle" caused by AI has been observed yet, so we need to remain cautious about its future impact.
In conclusion, the effects of AI on the economy are still in their early stages. It is neither a savior nor a disaster, and more time is needed to fully understand its true role.