• Goldman Sachs (GS) identifies labor exposure and low barriers to entry as key metrics for assessing AI disruption risk in stocks.
  • High-risk firms, particularly in software and services with automatable work, have underperformed, while protected ones with heavy physical assets have outperformed.
  • Markets are currently pricing in AI risks more than potential upsides, with recent volatility highlighting investor caution.

Goldman Sachs' latest analysis provides a framework for investors navigating the turbulent AI landscape, pinpointing two critical factors that separate winners from losers in the stock market. According to the firm's research, labor exposure—measured by high labor costs in automatable sectors like software and services—and low barriers to entry, such as heavy physical assets, serve as key metrics for gauging AI disruption risk. This comes as AI capital spending by hyperscalers is projected to reach $527 billion in 2026, up from prior estimates of $465 billion, driven by third-quarter earnings revisions.

Recent market movements underscore the urgency of this analysis. Software and services stocks, hit hard by AI disruption fears, have sold off significantly, with Salesforce (CRM) down 38% yearly, according to people familiar with the matter. In contrast, AI platform stocks like database providers are outperforming amid rising corporate adoption. One source close to the discussions noted, "Investors are rotating away from debt-funded infrastructure toward revenue-linked cloud platforms, but the real story is how labor exposure is driving divergence." Efforts to reach Goldman Sachs for additional comment were not immediately successful.

Without a clear strategy to assess these risks, companies in vulnerable sectors could face prolonged underperformance. The broader economic context adds pressure: AI-driven labor displacement could slightly raise U.S. unemployment in 2026, with estimates suggesting up to a 0.3% increase under faster adoption scenarios. This has sparked what some are calling the "Great AI Scare of 2026," with surveys like Harvard Business Review highlighting widespread fears among stakeholders, including laid-off employees and investors in laggard stocks.

Looking ahead, the focus is shifting from infrastructure gains—semiconductors and data centers are up 44% year-to-date but face earnings risks—to productivity beneficiaries in software and services. Goldman predicts outperformance for low-risk firms with physical asset-heavy operations, while experts see a potential "productivity breakout" in enterprise AI. As one analyst put it, "The AI trade has evolved from unified rallies to dispersion, and now it's all about who's protected and who's exposed." This analysis aligns with J.P. Morgan (JPM)'s 2026 outlook, which flags AI risks and opportunities across industries, though without direct government policies influencing the current dynamics.

In related developments, Goldman Sachs has expanded to 13,000 engineers for its AI transformation, signaling a long-term commitment. Similar patterns are emerging elsewhere: Salesforce's Agent Force shows 330% year-over-year growth in annual recurring revenue despite broader pullbacks, and ServiceNow (NOW)'s Now Assist has achieved over $600 million in annual contract value. As the market grapples with these shifts, the key takeaway is that AI will create clear winners and losers, but for now, investors are hedging against the downsides more than betting on the upside.