Man and Machine

Key Takeaways

• Alphabet’s employee compensation flat; Meta’s rose 34% (2022‑2024)
• ~30% of gross profit goes to employees, machines, shareholders each today
• By 2026, 55‑60% of gross profit will fund AI compute
• Pichai spends an hour weekly tracking TPU and GPU usage
• Compute, not cash, is the primary scarce resource for Alphabet

Pulse Analysis

Capital allocation in traditional businesses is straightforward: cash flow funds capex, acquisitions, dividends, buybacks, or debt repayment. Tech giants like Alphabet and Meta, however, invest heavily in operating expenses that appear on the income statement, especially AI‑related compute. Evaluating these firms at the gross‑profit level reveals where real value creation occurs, since much of the benefit from AI spend materializes years later, making return variance higher than in classic capex projects.

In a recent Cheeky Pint interview, Sundar Pichai described how compute—specifically TPUs and GPUs—has become the limiting factor for Alphabet’s growth. He explained that he dedicates a focused hour each week to monitor compute allocation across teams, treating it like a budget line item. This granular oversight reflects a shift toward treating AI models as strategic assets, where early‑stage investments are judged by long‑term option value and total addressable market rather than immediate cash returns. The approach mirrors venture‑style capital discipline, prioritizing projects that can unlock outsized returns over the next five to ten years.

Quantitatively, both Alphabet and Meta currently allocate roughly one‑third of gross profit each to employees, machines, and shareholders. The mix is evolving: employee and shareholder shares are shrinking while machine spend is accelerating, projected to consume 55‑60% of gross profit by 2026. For investors, this signals that future earnings will be increasingly tied to AI infrastructure efficiency and the companies’ ability to monetize compute. Understanding this allocation trajectory is essential for assessing margin sustainability, ROIC prospects, and the competitive dynamics of the AI‑driven economy.