Google Invests $20 Billion in AI Infrastructure This Year as Revenue Lags Spending

According to Beating monitoring, Google will invest around $20 billion in AI data centers and devices this year. DeepMind Chief Strategy Officer Jasjeet Sekhon acknowledged that current AI revenue cannot support this spending level, creating a potential gap where capital is deployed ahead of revenue generation.

Sekhon cited Recursive Self-Improvement (RSI)—using AI to research and develop next-generation AI while continuously refining training methods—as the core long-term logic for the massive investment. However, true RSI does not yet exist; current systems can only optimize algorithms and training processes. Sekhon expects full RSI implementation within two years, meaning Google is using a technology goal that has not yet materialized to justify today's $20 billion spending.