Artificial intelligence investment is accelerating rapidly across corporate America, but Goldman Sachs says the technology has yet to produce a clearly measurable earnings advantage for most S&P 500 companies.
The latest earnings season was nevertheless strong. S&P 500 earnings per share advanced 31% year-over-year in the second quarter after stripping out exceptional income associated with private investment stakes. AI infrastructure companies, including hyperscalers, contributed around half of that growth as their combined earnings increased 54% from the previous year.
Corporate profit growth was also relatively broad. Once the energy sector and its benefit from higher oil prices are removed, the median company in the S&P 500 still generated a 14% year-over-year increase in earnings.
The disconnect, according to Goldman strategists led by Ben Snider, is that companies are still struggling to demonstrate how deploying AI is feeding through to their bottom lines. Only 11% of S&P 500 businesses provided quantified productivity improvements for individual AI use cases during their latest earnings calls, including areas such as coding and customer service.
An even smaller 2% of companies quantified AI’s direct effect on earnings, leaving that proportion unchanged from the first quarter.
“Q2 results showed a small and statistically insignificant difference in earnings growth between the companies quantifying AI productivity gains this quarter and other S&P 500 companies,” the strategists wrote.
The potential for a more visible earnings contribution is increasing, however, as businesses rapidly expand their AI budgets. According to the Ramp AI Index, median monthly AI spending per employee more than doubled from $5 in January to $12 in July. The increase has been much greater among the highest-spending companies, with businesses in the top decile lifting monthly expenditure from $240 to $650 per employee over the same period.
For now, those costs remain manageable relative to corporate revenues. Goldman estimates AI inference spending amounts to less than 0.5% of S&P 500 sales. Its IT Spending Survey also indicates that roughly two-thirds of businesses are funding AI initiatives by redirecting money already allocated elsewhere.
Software budgets are the most common source of those funds, accounting for 18% of reallocations, while labour represents 11%.
That shift has not yet translated into widespread pressure on the software industry. Although individual companies such as Starbucks are developing internal AI products capable of replacing external software, broader software revenue growth has strengthened modestly over recent quarters. Software stocks have also rebounded after previously coming under pressure.
For investors, Goldman says the uncertain distribution of AI’s eventual financial benefits is influencing market positioning. AI infrastructure businesses offering clear and immediate earnings growth have attracted investor demand, while the market has been less willing to make aggressive bets on which companies will ultimately convert AI-driven productivity improvements into lasting earnings growth.
Leave a Reply