The sustainability of debt-funded artificial intelligence data centre capital spending came under renewed investor scrutiny on June 23, 2026, as the combination of Alphabet's research leadership departures, rising Fed rates, and Korean memory contagion exposed the leverage embedded in the AI infrastructure trade.

Key Highlights

  • Investors on June 23 questioned whether hyperscaler AI capital spending, partially funded by debt issuance, could be sustained at current rates given rising interest rates and uncertain monetisation timelines.
  • The AI infrastructure trade had produced extraordinary returns across semiconductor, data centre equipment, optical networking, and energy storage names in the months preceding June 23.
  • Leveraged positioning through instruments including 3x semiconductor ETFs had concentrated risk in AI-adjacent equities, amplifying the June 23 selloff beyond what fundamental reassessment would have produced alone.
  • Alphabet's loss of two senior AI scientists served as a catalyst for broader questions about whether the research talent driving AI infrastructure investment could be retained at the pace required.

 

The sustainability of hyperscaler artificial intelligence capital spending emerged as the dominant investor concern on June 23, 2026, as a combination of macro pressures exposed the leverage embedded in the AI infrastructure equity trade and produced sharp declines across the entire AI supply chain.

The major US cloud providers had been investing at an extraordinary pace in AI data centre infrastructure, building out GPU clusters, high-bandwidth memory systems, optical networking fabrics, and power infrastructure at a scale unprecedented in the history of technology capital expenditure. Much of this investment had been partially debt-financed, with cloud providers issuing bonds and commercial paper to fund the buildout ahead of revenue generation from AI services.

Rising interest rates under Federal Reserve Chair Kevin Warsh increased the cost of that debt financing and raised the discount rate applied to future AI service revenues, creating a dual headwind for the investment case. At the same time, the monetisation timeline for AI services remained uncertain, with consumer and enterprise AI product adoption varying significantly across use cases.

On June 23, reports that Alphabet had lost two senior AI scientists provided the catalyst for investors to act on these concerns. The AI infrastructure trade had become so widely owned and so heavily leveraged through instruments including 3x semiconductor ETFs that forced rebalancing on a risk-off day produced extreme single-session moves across all AI-adjacent names.

The selloff affected every segment of the AI supply chain: memory chips, semiconductor equipment, custom ASIC designers, optical networking, data centre power and cooling, and clean energy storage names associated with data centre power infrastructure all declined sharply and simultaneously, illustrating the degree of cross-asset correlation that had developed within the AI infrastructure investment theme.