The US stock market has shown little upward momentum in recent months, with the S&P 500 moving sideways amid heightened volatility in technology shares. This shift emerged notably in May when the "Magnificent Seven" tech giants—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla—lost their long-held leadership role in driving market gains. In the weeks that followed, semiconductor stocks such as Micron and Broadcom briefly took the lead but have since also waned. Adding to the market turbulence, an AI-focused hedge fund, Situational Awareness, recently collapsed, highlighting the risks and uncertainty engulfing technology sectors.

Despite concerns of irrational exuberance and elevated valuations, current market behavior appears to reflect a deeper reassessment rather than mere panic. Investors are grappling with the evolving competitive dynamics presented by the artificial intelligence (AI) industry and its broader impact on the economy. Central to this reconsideration is the question of what the competitive structure of AI will look like, especially as the technology integrates across multiple sectors.

The prevailing framework in recent decades has been what economist W. Brian Arthur termed the "increasing returns era." Unlike traditional manufacturing industries, knowledge-based or digital industries—such as those dominated by the major tech firms—benefit from high upfront development costs but almost negligible marginal costs for additional units. These sectors also enjoy strong network effects and significant switching costs, allowing dominant companies to maintain substantial competitive moats. Research by Hendrik Bessembinder underscores this concentration, showing that nearly half of the net wealth creation in US stock markets over the past century has come from just 46 public companies, many of which are knowledge-based enterprises.

However, AI threatens to disrupt this paradigm by commoditizing knowledge itself. There are two primary concerns. First, AI tools could render certain knowledge-based businesses obsolete, a phenomenon that has unsettled the shares of leading software-as-a-service companies in what some observers call the “SaaSpocalypse.” More fundamentally, frontier AI models may become highly similar and interchangeable, presenting limited opportunities for firms to build defensible intellectual property or leverage significant switching costs. Unlike software or internet services, ongoing operational expenses related to hardware depreciation and energy consumption mean that unit costs may not decline dramatically. This scenario could transform AI into a widely accessible utility akin to electricity—highly beneficial for the broader economy but potentially disappointing for investors seeking outsized returns. Notably, utilities are historically absent from lists of top wealth creators.

Markets appear to be taking seriously the possibility that future AI-driven industries may offer lower profitability than the traditional tech giants. This partly explains the cautious investor reaction when affordable AI models or hardware emerge from Chinese competitors. It also sheds light on why semiconductor companies have struggled to maintain market leadership after the decline of the Magnificent Seven, as their margins depend heavily on the profitability of AI services offered by the tech leaders.

As a result, investors are revising their expectations, acknowledging increased risks for major technology firms and adjusting valuations accordingly. There is also growing recognition of the potential renewed importance of more traditional competitive advantages such as brand strength, distribution networks, and manufacturing expertise. This evolving landscape signals that the era dominated by a handful of knowledge giants may be giving way to a more diverse and competitive economic environment.