The AI trade is hitting a wall of reality. After months of unchecked optimism, investors are suddenly questioning whether the hundreds of billions of dollars flowing into data centers will ever actually materialize as high-margin revenue. This shift in sentiment triggered a fear-driven semiconductor sell-off that erased $2.7 trillion in market capitalization from some of the year’s biggest winners.
The $700 Billion CapEx Question
Wall Street is no longer taking “AI potential” as a proxy for profit. The four largest hyperscalers—Alphabet, Amazon, Meta Platforms, and Microsoft—are on track to spend more than $700 billion on capital expenditures (CapEx) this year. However, the cost of this build-out is rising.
This inflation is trickling down to consumers. Apple has already announced price hikes for Macs and iPads, citing a memory chip shortage. Similarly, Microsoft increased XBOX console prices, and Nintendo has a Switch 2 price increase set for Sept. 1. The concern is that these skyrocketing input costs are creating a rift between what hyperscalers consume and what memory suppliers produce.
South Korean Margin Calls and the Kospi Crash
While the U.S. market felt a dip, South Korea experienced a collapse. The Kospi index plunged 27% after hitting a record high last month, largely due to a “trading mania” around chipmakers Samsung Electronics and SK Hynix. This rally was fueled by leveraged ETFs and high-risk borrowing.
The bubble burst when the market turned, triggering a massive chain reaction of forced liquidations. According to Tom Lee, cofounder of Fundstrat Global Advisors, 1.2 million brokerage accounts faced margin calls, representing up to 10% of all accounts in Korea. As traders sold holdings to cover their debts, the fallout spilled into global markets.
The impact was felt acutely in Japan. On July 24, the Nikkei share average ended more than 2% lower, falling 2.73% to close at 64,611.15. The decline was exacerbated by a 7% drop in Alphabet shares overnight, which rekindled fears about the sustainability of AI infrastructure spending.
Software Disruption: The Anthropic Effect
It isn’t just the hardware makers feeling the heat. Traditional software companies are being punished as AI startups move from “chatbots” to “agents” that can actually do the work. The release of new tools by AI startup Anthropic, designed to automate tasks in legal, financial research, and data services, sparked fears that established software business models are now obsolete.
The resulting panic led to a massive loss of value. A group of 164 stocks across software, financial services, and asset management shed US$611 billion (S$776 billion) in market value in a single week.
- Thomson Reuters (Canada): Plunged 20% last week, its steepest fall ever.
The speed of this evolution is dizzying.
Daniel Newman noted that things are shipping out weekly and daily, and that the blast radius of companies that could be impacted by AI is going to grow daily.
The Bull Case for Memory Chips
Whether this is a rolling correction that is super healthy or the start of a permanent disruption depends on the next wave of earnings. For now, the market remains caught between the raw growth of memory chip makers and the looming threat of AI-driven obsolescence for the software companies that once anchored the S&P 500.
Keep reading
Discover more from Archyworldys
Subscribe to get the latest posts sent to your email.