AI Jitters and Rising Yields: Looking Beyond the Headlines

AI Jitters and Rising Yields: Looking Beyond the Headlines

Stocks slipped last week, with the S&P 500 down 0.6%, the Dow down 0.4%, and the Nasdaq Composite down 2.1%, as semiconductor stocks led the decline, with the sector’s index falling 4.3%. Rising Middle East tensions pushed oil prices higher and reignited inflation concerns, while renewed scrutiny of AI infrastructure spending weighed on technology shares following Alphabet and Tesla’s earnings reports.

Treasury Yields Climb Continues

The 10-year Treasury yield climbed above 4.7% on Thursday, its highest level since January 2025, before easing slightly to close the week near 4.69%. Brent crude touched $100 a barrel intraweek before pulling back to settle near $97 on Friday, and fell further, below $86, in Monday morning trading on hopes that a pause in the Iran conflict could lead to renewed peace talks.

AI Spending Questions

Much of last week’s technology weakness traces back to a familiar question investors are asking as mega-cap earnings season continues: will the hundreds of billions being spent on AI infrastructure ultimately generate sufficient returns? It’s a good moment to step back and remember that “AI” is not a single investment. Model providers such as OpenAI, Anthropic, and Google represent only one piece of a broader supply chain that also includes semiconductor hardware, data centers, and the software providers building AI into everyday business tools.

AI Chip Demand

At the foundation is hardware, chiefly GPUs and memory chips, supporting two distinct phases of demand: training, the resource-intensive process of building large language models over weeks or months, and inference, the ongoing computational cost of running these models each time someone submits a prompt. Both have driven strong demand and pricing for semiconductors, and both help explain why chip stocks move so sharply on any shift in the AI spending narrative.

Data Center Surge

That hardware has to be housed somewhere, and data center construction spending has accelerated sharply since ChatGPT’s late-2022 launch, now surpassing all other categories of office construction. A key debate among investors is the “Jevons paradox”: whether more efficient AI models reduce computing demand, or whether cheaper, more capable technology instead broadens adoption and creates new use cases, as happened historically with electricity and computing.

High Expectations Ahead

Valuations have risen alongside AI enthusiasm. The Information Technology sector trades at roughly 21.4 times earnings, elevated relative to its own history and the broader market, with Communication Services and Consumer Discretionary, which also house large technology companies, similarly stretched. These valuations partly reflect genuinely strong earnings growth tied to AI demand, though they leave less room for disappointment, as last week’s chip-driven pullback showed.

Looking Back

History is a useful check on timing: investor enthusiasm for internet stocks in the late 1990s proved directionally correct but took decades to fully play out. With the Fed’s policy decision due Wednesday and Meta, Microsoft, Apple, and Amazon all reporting earnings this week, further volatility in AI-related names is likely. The takeaway for investors is to maintain a broad view across the full AI supply chain, alongside other attractively valued areas of the market, rather than reacting to swings in a handful of headline names.

As Always

I’d like to leave you with the final line we’ve used since we started these commentaries back at the very height of market volatility in March 2020. Always remember that we create financial/investment plans not for the easy times, but to prepare for the tough ones.