AI Spend, Model Convergence, and the Widening Gap in Tech Stocks

The AI Spend Reckoning

• 💰 Investors are punishing big AI spenders that lack clear returns, leading to increased capex and talent costs that squeeze margins.
• 📈 Hyperscalers have issued $121 billion in new debt, four times the 5-year average, indicating that the AI buildout is outrunning internally generated cash.
• 📉 Companies like Amazon and Meta are in the "lagard camp" due to massive investments in data centers and AI hires without significant market share in key AI segments.

Model Convergence and Distribution's Edge

• 🧩 As the gap between AI models narrows, distribution becomes more critical than the model itself.
• 🚀 Alphabet is leading due to its ownership of the full stack: chips, cloud, distribution, and top-performing AI models, even without full credit for its Waymo venture.
• 🚗 Tesla's high PE ratio is contrasted with Waymo's lead in autonomous driving, highlighting the market's speculative nature versus current technological advantage.
• 📱 Apple, with its vast distribution of 2.3 billion active devices, is positioned to be a competitive force in 2026, even with lower infrastructure spend.

Challenges with AI Partnerships and Debt

• ⚠️ Oracle is a poster child for problematic debt, with a significant backlog tied to OpenAI's $300 billion commitment, raising concerns about payment before capacity is online.
• 🏦 Microsoft also faces similar concerns with OpenAI's $300 billion commitment to Azure, as OpenAI commits to over $1.4 trillion in compute without secured revenue or funding for these bills.
• 🤝 The reliance on single customers and large, unfulfilled commitments creates significant financial risk for companies involved in the AI infrastructure buildout.