We broke down the eye-popping AI spending for 4 Big Tech firms — and their plans to go even harder next year (www.businessinsider.com)

Big Tech just reopened the spending floodgates on AI: Amazon, Microsoft, Google and Meta reported record capital expenditures this quarter and signaled even bigger capex ramps into 2026. Microsoft led with nearly $34.9B (a 74% YoY jump), roughly half of which went to GPUs/CPUs and $11.1B to data‑center leasing. Amazon spent $34.2B this quarter and is targeting roughly $125B for the year (up from $83B in 2024), expanding Project Rainier with an $11B Indiana campus and $5B planned in South Korea. Google spent $24B—about 60% on chips/servers and 40% on networking and facilities—and raised annual capex guidance to $91–93B. Meta spent $19.4B, has boosted its 2025 capex range and is planning hefty infrastructure increases next year, funded in part by $29B in project financing for a Louisiana data‑center build. Microsoft also joined a consortium to buy Aligned Data Centers for $40B, underscoring consolidation in the market. The significance is twofold: technically, this capex funds the scale needed for large model training and high‑throughput inference (more GPUs, custom servers, network and real estate), accelerating product and cloud AI capabilities; financially, it raises stakes for monetization—investors reward companies showing clear AI returns (Google) and are growing impatient with those that haven’t yet (Meta). The rush creates downstream winners (chip makers, colo operators) but also risks overcapacity and an “AI bubble” if revenue doesn’t follow the escalating cost base.