AI Slowdown Fears Hit Chip Stocks, Boost Hyperscalers

Artificial intelligence's existential risk debate triggered a significant market divergence on Monday, impacting technology stocks. Chipmakers experienced a notable downturn, with Nvidia falling over 3% and Intel, AMD, and Marvell dropping between 5% and 6%. This decline dragged the Philadelphia Semiconductor Index down by nearly 6%. Companies heavily reliant on data center construction, such as Neoclouds, also faced downward pressure. In contrast, major cloud service providers, identified as hyperscalers and significant investors in AI infrastructure, saw their stock prices rise. Alphabet increased by almost 2%, Microsoft gained 1.6%, and Meta added approximately 1.4%. Amazon experienced a slight dip of about 1.6% but demonstrated resilience compared to the semiconductor sector.

This market reaction reflects investor strategies concerning a potential slowdown in the frontier AI race, a risk that had previously been a quiet consideration in Wall Street's risk models. The catalyst for this shift appears to be recent calls for enhanced AI safety measures and a potential deceleration in model development. On Saturday, Anthropic CEO Dario Amodei advocated for stronger safety protocols and a slowdown in AI progress, citing alarming demonstrations of AI's cyber capabilities. OpenAI CEO Sam Altman and other prominent AI leaders publicly supported these sentiments. Some proposals even suggested government intervention to enforce such a slowdown.

Gil Luria, head of technology research at D.A. Davidson, provided an analysis of the potential market impact. Luria argued that a slowdown in AI development would disproportionately affect companies supplying the underlying infrastructure, often referred to as the "picks and shovels" of the AI trade, such as Nvidia and CoreWeave. Conversely, hyperscalers like Microsoft, Amazon, and Google, who are the primary consumers of this infrastructure, would be less impacted. Luria explained that if AI continues its exponential improvement, these hyperscalers will maintain their substantial investments in building data centers to meet escalating demand. However, if the pace of AI progress decelerates, these companies have the flexibility to halt further capacity expansion and instead focus on optimizing and generating returns from their existing infrastructure.

Luria elaborated on this scenario, predicting that "They’ll just all stop building data centers and just digest what they have." This strategic shift could lead to continued revenue and profit growth for hyperscalers, even as their capital expenditures decrease. Such a dynamic would result in higher free cash flow. Luria drew a parallel to Amazon's post-pandemic adjustment, where the e-commerce giant scaled back its warehouse construction following a period of rapid expansion.