Speculation about a slowdown in artificial intelligence development put pressure on equity markets on Monday. Anthropic co-founder and CEO Dario Amodei published an essay over the weekend advocating for a reduced pace of AI model development, amid an accumulation of public debate and media coverage around AI safety. Sam Altman and Elon Musk have each publicly backed calls for a slower pace of AI development and greater regulatory oversight of the technology.
Notwithstanding public statements from both the US President and China's Foreign Ministry opposing constraints on AI development, equity markets in the United States recorded notable declines. Semiconductor and other hardware equities recorded the steepest declines as markets adjusted to the possibility that a slower pace of AI model advancement could weigh on the broader AI infrastructure spending cycle that had been a principal driver of market gains over the prior year.
James Learmonth, co-CIO and Portfolio Manager at Harvest ETFs, noted that while calls for reduced development speed and greater oversight of foundational AI models had only just surfaced, sentiment around AI had already been shifting over the preceding months, with public resistance to data centre expansion and rising treasury yields both contributing to a more cautious outlook on investment costs. Learmonth added that the latest development had reinforced pre-existing market unease rather than introducing an entirely new concern, and that the overall reaction reflected a short-term response to news driven primarily by sentiment.
No shift in the underlying fundamentals is anticipated. "Nobody is announcing plans to change their spending, their builds, or their operations," said Elliot Johnson, CIO at Evolve ETFs, though he noted that investor nervousness was evident in market behaviour.
With both nations having publicly stated their commitment to advancing artificial intelligence, and one having framed the pursuit of superintelligence as a direct technological rivalry with the other, momentum behind the development of such systems is not expected to slow.
Learmonth observed that Anthropic has discussed independently deploying third-party evaluators regardless of whether governments establish an AI regulatory framework, though the company also faces rivalry from OpenAI, XAI, and other providers — pressure that may intensify as it moves toward an IPO. According to Johnson, firms may have additional reasons to decelerate AI model development, among them the need to establish viable unit economics within the sector. He observed that a more sceptical interpretation would be that these firms are pushing for regulatory frameworks primarily to entrench their existing market dominance.
Johnson and Learmonth both took the position that a deceleration in AI development would be unlikely to alter the trajectory of infrastructure investment over the near to medium term. Johnson raised a hypothetical scenario: if AI development were to plateau at its current state, would adoption rates rise, fall, or remain unchanged over the next five years? He expressed the view that user numbers will grow regardless of whether AI models are updated.
Purchasing of GPUs, DRAM chips, and related components is expected to persist across hardware and semiconductor sectors. According to Learmonth, the decline in hardware equities reflects a shift in the timing of investor expectations around AI data centre expansion rather than any fundamental change in market consensus. Growth in AI services is expected to persist, though some investors anticipate the pace may ease in the near term.
Analysts have suggested that a modest deceleration in demand may ease existing supply constraints, potentially strengthening the pricing position of certain hardware manufacturers. Learmonth further noted that doubts about the long-term viability of AI expansion were already reflected in the valuations of certain stocks, and that financial incentives to develop AI infrastructure further would persist even if new regulations were introduced.
Johnson argued that AI governance falls outside the domain that specialists can readily address. Instances of AI operating outside intended parameters often arise from unforeseen and unplanned system behaviours, posing considerable difficulties for those seeking to regulate and enforce compliance.
Johnson and Learmonth both noted that portfolio diversification has continued to demonstrate its worth amid the prevailing market conditions. Learmonth called for broad exposure across the technology sector, spanning both hardware equities and software holdings that had recovered some ground following discussions around an economic slowdown, while Johnson noted that no reports had surfaced of companies planning to reduce capital expenditure. No indication has emerged that consumers are expected to curtail their use of these goods and services.
The equity market rally over the past 18 months has been driven primarily by margin expansion rather than multiple expansion, reflecting sustained quarterly earnings growth from large technology companies — a pattern analysts generally view as a sound foundation for the advance.
