AI Bubble Fears Intensify as Economists Warn of Hidden Leverage and Political Backlash

Key points

• Dario Perkins of Global Data.TS Lombard argues that AI revenue growth is currently driven by capital expenditure recycling within the ecosystem, not by external business or consumer adoption.
• Perkins estimates that debt across the AI ecosystem could reach $2 trillion by the end of 2027, creating vulnerability for smaller cloud providers and infrastructure suppliers.
• Bank of America strategist Sebastian Raedler notes that hyperscaler spending is projected to hit $1 trillion in the next 12 months, a level exceeding the 1870s railroad boom relative to U.S. GDP.
• The Philadelphia Semiconductor Index fell more than 11% last week, reflecting concerns over potential oversupply in chip manufacturing.
• Perkins highlights political backlash as a potential trigger, noting that public opposition to data centers and job displacement could turn AI into a toxic political issue.

Background

Wall Street has been grappling with growing concerns about an AI bubble, particularly following a recent selloff in chip stocks. The Philadelphia Semiconductor Index dropped over 11% last week before partially recovering. This volatility has prompted analysts to scrutinize the sustainability of massive infrastructure spending, which has surged from $330 billion in early 2025 to an expected $1 trillion in the next year.

How outlets are covering it

Business Insider highlights the risk of 'capex recycling,' where AI companies spend money on each other rather than generating revenue from external users, and warns of political backlash from the public. In contrast, TheStreet, via Bank of America, emphasizes historical parallels, noting that past capital cycles often ended in disappointment due to overcapacity. While Bank of America focuses on the cyclical nature of hardware and falling token prices, Business Insider points to structural weaknesses in debt levels among smaller firms and the potential for a single high-profile accident to trigger broader opposition. Some bank CEOs, such as Jane Fraser of Citigroup, remain bullish, arguing that compute demand remains high, while others like Charlie Scharf of Wells Fargo express caution regarding leverage.

Why it matters

If the AI boom relies heavily on internal spending rather than external revenue, a slowdown could trigger a cascade of defaults among leveraged smaller firms. Additionally, political opposition to data centers and job displacement could lead to regulatory or legislative hurdles that disrupt the industry's expansion, potentially causing a broader market correction beyond just tech stocks.

What to watch

Investors will monitor the Philadelphia Semiconductor Index for further volatility and watch for signs of overcapacity in chip manufacturing. There will also be increased scrutiny on the debt levels of smaller AI infrastructure providers and any political developments regarding data center approvals or AI regulation. Bank earnings calls will continue to provide insights into the sustainability of AI-related lending and spending.