Despite hopes of ushering in a new era of rapid growth and affluence, AI is not yet clearly visible in macro data, according to Torsten Slok, Apollo’s chief economics officer.
In a memo Saturday, he recalled economist Robert Solow’s quip from the 1980s, when PCs were transforming the economy, that “you can see the computer age everywhere except in productivity statistics.”
The same is true for AI today, Throck wrote, noting that data on employment, productivity, and inflation still show no signs of the new technology. There is also no evidence of AI at work in the margin and revenue forecasts of S&P 500 companies outside of the Magnificent 7.
“AI is everywhere except in incoming macroeconomic data,” he said.
To be sure, investors aren’t waiting for AI to upend business models, and their concerns have been draining the stock market lately.
As increasingly sophisticated chatbots are deployed, stocks related to wealth management, insurance brokerage, tax preparation, accounting services, professional data, legal research, trucking, and logistics are selling big.
Meanwhile, AI evangelists see incredible economic benefits. Anthropic CEO Dario Amodei said at the World Economic Forum last month that AI could boost GDP growth by 5% to 10%.
And xAI co-founder Elon Musk predicted that AI will create so much wealth that working in the not too distant future will become an option.
But Throck remains unconvinced.
“Maybe there is a J-curve effect of AI, where it takes time for AI to show up in macro data. Maybe there isn’t,” he wrote on Saturday.
It all depends on value creation through AI, Slok explained. So far, things have played out differently than the computer revolution of the 1980s.
Instead of early innovators enjoying exclusive pricing rights until they are eclipsed by competitors, fierce competition among large language model developers drives prices for end users closer to zero.
But from a macro perspective, the value created by AI comes from how it is used in the economy, rather than from specific products, Throck said. Economists have so far said in several studies that they do not foresee a major impact.
For example, in Penn Wharton’s budget model, AI would increase total factor productivity by only 0.1 to 0.2 percentage points per year, resulting in a cumulative 1.5% increase by 2035.
apollo global
“After three years of using ChatGPT, there is still no sign of AI in the data coming in. AI will probably augment the workforce in some areas rather than replace it in all areas,” Slok said.
Similarly, the Congressional Budget Office has taken a relatively conservative view, estimating that AI will increase total factor productivity by just 0.1 percentage point per year, ultimately boosting output by 1 percentage point by 2036.
However, this also coincided with the Department of Labor revising its 2025 employment growth statistics to just 181,000 jobs, down from the original 584,000 and 1.46 million jobs in 2024.
Given that the economy continued to expand at a healthy pace despite very low payroll growth last year, productivity should skyrocket and questions should arise about what impact, if any, AI had.
“The widespread adoption of generative AI applications currently in operation is expected to improve operational efficiency and organization of work, thereby providing a modest boost to TFP growth over the next decade,” the CBO said in its latest forecast.
