Paul Meeks on AI Debt, Infrastructure Spending, and Investment Opportunities

AI Debt vs. Late 90s Infrastructure Borrowing

• 💡 AI debt is not considered a problem akin to the late 90s/early 2000s when companies with weak balance sheets borrowed heavily for internet infrastructure.
• 📌 This time, relationships are more on an equity basis than debt, and the primary builders are hyperscalers with strong balance sheets and cash flows.
• ✅ The majority of AI infrastructure is being built by five hyperscalers who could easily borrow if needed, but most have not yet.

Depreciation Concerns and Investment Focus

• 🎭 Regarding depreciation concerns, the situation is described as potentially "somewhere in between The Tempest and much to do about nothing."
• 📈 While lengthening asset life can lower early depreciation and boost earnings, the focus for analysts is on revenue growth in the short term (2025-2027) as infrastructure is built, rather than immediate earnings growth.
• 🔍 The long-term ROI from AI investments is acknowledged, but immediate concerns are less pressing.

AI Infrastructure Trade Outlook

• 🚀 The AI infrastructure trade is seen as an opportunity for at least another year.
• 📊 CEOs indicate significant spending increases in 2026 on top of a high 2025 base, suggesting this trend is locked in.
• ⚠️ While growth may slow and expenditures plateau in 2027, a collapse is not expected.

Investment Opportunities in AI

• 🎯 Key companies of interest include Nvidia, CoreWeave, Nebius Group, and Applied Digital, all involved in AI technology assets.
• 📉 Despite CoreWeave's stock drop after a guidance miss, the company is seen as fundamentally sound, presenting a potential buying opportunity.
• 🤝 The interconnectedness of these companies with Nvidia is highlighted, especially with Nvidia's upcoming announcements.