Goldman Sachs strategist Ben Snider has pointed out that while companies are accelerating their investments in artificial intelligence (AI) at an unprecedented pace, the technology has yet to translate into substantial earnings improvements for most firms. In a report released on August 14, Goldman Sachs noted that during the second-quarter earnings season, only 2% of S&P 500 companies quantified AI's specific impact on earnings, while 11% reported measurable productivity gains in specific areas such as software coding and customer support. However, those companies that achieved efficiency improvements did not see earnings growth significantly outperform the broader market. According to Goldman Sachs, the median year-over-year earnings growth for these companies was 17%, compared to 14% for those that did not quantify AI's efficiency contributions. The bank noted that this gap is not statistically significant.
For investors, this finding helps explain the current market landscape: on one hand, AI infrastructure beneficiaries such as semiconductor manufacturers and cloud service providers continue to gain favor; on the other hand, companies promising future efficiency improvements are being met with widespread caution. Infrastructure spending has already delivered immediate revenue and profit growth, while the potential returns from companies using AI to enhance efficiency remain difficult to measure and may take several quarters to materialize.
Overall, the second-quarter earnings season was particularly strong. Goldman Sachs stated that, excluding non-recurring gains related to certain private equity investments, S&P 500 earnings per share rose 31% year-over-year. Among hyperscale companies and other beneficiaries of AI capital expenditures, earnings surged by 54%, contributing roughly half of the index's overall earnings growth. Despite this, growth momentum was not limited to large-cap tech stocks. The median earnings growth for S&P 500 components was 14%; excluding energy companies that benefited from rising oil prices, the growth rate for non-AI infrastructure companies was also 14%. This broader improvement could help alleviate concerns that earnings growth is overly reliant on a handful of tech giants. Nevertheless, the performance gap between infrastructure suppliers and AI application companies remains significant. Goldman Sachs noted that investors favor infrastructure stocks because their returns are immediate and relatively easy to track. In contrast, portfolios of companies that frequently mention AI productivity initiatives have roughly matched the broader S&P 500 index over the past few years.
Evidence suggests that AI's impact may become clearer in corporate earnings reports over the coming quarters. Goldman Sachs, citing the Ramp AI Index, showed that average monthly AI spending per employee has risen from $5 at the start of the year to $12 in July, while spending among the top 10% of companies surged from $240 to $650. During second-quarter earnings calls, approximately 7% of S&P 500 companies discussed AI deployment costs. Most companies indicated that related spending remains small, emphasized a prudent approach to investment, or stated that AI-driven benefits have already exceeded costs. Goldman Sachs estimates that AI inference costs currently account for less than 0.5% of revenue for S&P 500 companies. Its latest IT spending survey found that 89% of respondents allocate between 1% and 5% of their IT budgets to AI. It is worth noting that these estimates do not include all costs associated with AI deployment, such as staffing and technical infrastructure development.
Goldman Sachs' survey found that about two-thirds of companies are funding AI investments by reallocating resources from existing budgets, rather than relying entirely on new capital. Specifically, 35% of respondents said AI spending comes from new budgets; 18% fund it through cost-efficiency projects; another 18% reallocate from software budgets; 11% from labor costs; 10% from cloud service budgets; and 9% from data analytics spending. Goldman Sachs believes that structural shifts within IT budgets are more likely to redistribute profits among different companies than to significantly alter overall S&P 500 earnings levels. However, if labor costs are substantially reduced, it could have broader economic implications. For now, the impact on the labor market has been concentrated in marketing, graphic design, customer service, and some technical roles, while job creation in data center construction has partially offset these losses. Goldman Sachs economists expect AI to eventually replace some labor, but they view this impact as temporary and smaller than many investors anticipate.
Market concerns that customers might use AI to develop their own applications, reducing reliance on external software vendors, have not yet materialized on a wide scale. In Goldman Sachs' IT survey, only 17% of respondents plan to increase internal software development and reduce purchases of off-the-shelf software. The median annual recurring revenue (ARR) growth rate for software companies covered by Goldman Sachs rose from 18% in the fourth quarter of 2025 to 22% in the first quarter of this year, accelerating further to 23% in the second quarter. This does not mean individual suppliers can rest easy. Goldman Sachs noted reports that Starbucks (SBUX.US) is developing internal AI tools to replace some software provided by companies like Microsoft (MSFT.US) and IBM (IBM.US). However, industry-wide data does not yet show a widespread deterioration trend.
Goldman Sachs suggests that companies with high labor costs and significant potential for automation in many roles may ultimately benefit most from AI. Currently, labor costs account for about 12% of total S&P 500 revenue, or roughly $2.1 trillion annually. Sector differences are significant: labor costs make up 21% of revenue in industrials, 16% in information technology, and just 5% in energy. Goldman Sachs screened Russell 1000 components for companies with high labor costs, strong potential for AI-driven automation, and management discussions of AI-related efficiency improvements in earnings calls. The list includes: Costar Group (CSGP.US), Dollar Tree (DLTR.US), eBay (EBAY.US), Arthur J. Gallagher (AJG.US), Axon Enterprise (AXON.US), The Trade Desk (TTD.US), Airbnb (ABNB.US), Boeing (BA.US), Lockheed Martin (LMT.US), Charles Schwab (SCHW.US), and Morgan Stanley (MS.US). However, Goldman Sachs cautioned that this screening list does not imply these companies have already achieved significant AI-driven cost reductions or efficiency gains—in fact, the earnings data for these potential beneficiaries does not yet show substantial improvement. For now, the AI investment thesis remains clearly divided: on one side are infrastructure providers, whose profits are already visible; on the other are technology deployers, whose financial returns remain more in the realm of expectations.
Comments