Big Tech’s capital spending has exploded since ChatGPT’s release in 2022, marking the start of an AI capex supercycle. Microsoft and Amazon are driving the largest increases, with Meta sharply scaling and Alphabet steady but smaller. Annual capex across the four could exceed $500bn by 2030, on par with national infrastructure projects. The surge reflects an arms race for compute, data centers, and networking capacity, shifting AI from a software play to capital-heavy infrastructure economics. While this build-out underpins AI’s ubiquity, it also raises bubble risks if monetization lags demand, echoing past tech cycles of boom and bust. Yet whatever happens here this is infrastructure that will serve AI adoption in the long term.