In the Broader Economy, AI Costs and Concerns Far Exceed Actual Impact

Silicon Valley's AI hype machine remains in overdrive despite growing concerns from surprisingly cautionary reports by such traditional fellow-travelers as Goldman Sachs. 

But the purveyors of concerns about excess costs and underperformance are growing rather than abating and the weight of evidence suggests that the corporate world is treading lightly and slowing when it comes to AI's exaggerated promises. The result suggests that AI's economic impact remains a ways from becoming measurable, let alone impactful. JL

Jon Sindreu reports in the Wall Street Journal:

AI’s knack for easy tasks has led to exaggerated predictions of its power to enhance productivity in hard jobs. Data from the Census Bureau show only a small percentage of U.S. companies outside of the information sector are looking to make use of AI. Developers of AI applications have made eye-watering capital expenditures, which are responsible for the profit surge in the rest of the ecosystem. (But) the profitability of implementation-focused AI companies is more important than that of their “upstream” counterparts. Big spending plans seem to be part of the reason markets reacted negatively to Alphabet’s results. AI firms won’t be able to goose one another’s profits forever.

Whatever you think of it, generative artificial intelligence has proven adept at something: generating profits for other AI projects. That is risky.

With more than 90% of companies in the S&P 500 having already reported second-quarter results, earnings are on track to rise at their fastest rate since 2021. AI-related corporations have significantly beaten analysts’ expectations, yet have lost 5% of their market value since the end of June. Part of this is because of an unwinding of trades using borrowed money, from which the market is already recovering.

 

Digging a bit further into the broad moniker of “AI firms,” however, might also help explain the market’s recent concerns.

As in most industries, the AI supply chain involves both “downstream” and “upstream” companies. Research shop Bespoke Investment Group has conveniently created equity baskets for each. The former group involves “AI implementation” firms that develop tools such as the large language models popularized by OpenAI’s ChatGPT since the end of 2022, or run products that can incorporate them. This includes most of the Magnificent Seven, as well as other members of the S&P 500 such as IBM, Adobe and Salesforce.

Higher up the supply chain, by contrast, are the “AI infrastructure” providers, which sell chips, data centers and training software. The undisputed leader is Nvidia, which has seen its sales triple in a year, but it includes other semiconductor firms such as Intel and Qualcomm, database developer Oracle and owners of data centers Equinix and Digital Realty. 

Crucially, these latter companies are the ones that have delivered most of the upside for investors, by posting profit margins that are far above what analysts expected a year ago. In the second quarter, and pending Nvidia’s results on Aug. 28, upstream AI members of the S&P 500 are set to have delivered a 50% annual increase in earnings. For the remainder of 2024, they will be increasingly responsible for the profit growth that Wall Street expects from the stock market—even accounting for Intel’s travails.

To be sure, the lines between the two groups can be blurry, particularly when it comes to giants such as Amazon.com and Alphabet, which also provide key AI infrastructure: Their cloud-computing arms are responsible for turning these companies into the early winners of the AI craze last year and reported breakneck growth during this latest earnings season. 

The important point, though, is that it is their role as ultimate developers of AI applications that have led them to make eye-watering capital expenditures, which are responsible for the profit surge in the rest of the ecosystem.

 

Annual earnings growth for these implementation-focused AI firms is likely to have slowed to around 22% in the second quarter, from a peak of roughly 50% last year. During the fourth, it is expected to come in at 8%.

As the path for monetizing this technology gets longer and harder, the benefits seem to be increasingly accruing to companies higher up in the supply chain. Meta Platforms Chief Executive Mark Zuckerberg recently said the company’s coming Llama 4 language model will require 10 times as much computing power to train as its predecessor.

Were it not for AI, revenues for semiconductor firms would probably have fallen during the second quarter, rather than rise 18%, according to S&P Global.

 

This isn’t as strange as it might seem. For the economy overall, the spending that makes up corporate profits must come from somewhere. As Polish economist Michał Kalecki noted, if trade and the government’s budget are balanced and consumers don’t spend more than their wages, it is corporations through capital expenditure that ultimately create their own earnings. In the case of the U.S., of course, wide budget deficits have flattered profits. But such macroeconomic tailwinds are now easing, and private investment must take over.

 

The problem is that investment in AI isn’t widespread. Capital expenditures among the non-AI bulk of the S&P 500 are up, but mostly follow pre-Covid trends. Private fixed investment across the economy is growing mostly thanks to manufacturing, which has been aided by the Inflation Reduction Act and the Chips Act. Some of this is related to semiconductors, but there has been no surge in overall tech and software investment, official data show. Nor has AI made a dent in other developed economies.

In a recent working paper, Massachusetts Institute of Technology economist Daron Acemoglu argued that AI’s knack for easy tasks has led to exaggerated predictions of its power to enhance productivity in hard jobs. Indeed, data from the Census Bureau show that only a small percentage of U.S. companies outside of the information and knowledge sectors are looking to make use of AI.

Disaster isn’t necessarily around the corner. After all, non-AI firms’ earnings growth finally seems to be catching up. Also, it is to be expected that widespread adoption of a new technology will take time. 

Nevertheless, investors should remember that the profitability of implementation-focused AI companies is ultimately more important than that of their “upstream” counterparts. Big spending plans seem to be part of the reason markets reacted negatively to Alphabet’s results a few weeks back. AI firms won’t be able to simply goose one another’s profits forever.