Google Q2 Capital expenditures doubled and broke records: US$44.9 billion was spent on AI infrastructure, and cloud business profit margins almost doubled. Google’s computing power bill is climbing at an alarming rate.

Google Q2 Capital expenditures doubled and broke records: US$44.9 billion was spent on AI infrastructure, and cloud business profit margins almost doubled. Google’s computing power bill is climbing at an alarming rate. Alphabet’s second quarter financial report shows that capital expenditures surged 100% year-on-year to a record high of US$44.92 billion. At this pace, annualized, the full-year expenditure has approached US$180 billion. The real story is the money-making machine behind the money that turns faster and faster. In the second quarter, Google's revenue increased by 24% year-on-year to US$119.8 billion, higher than analysts' estimates of approximately US$116.9 billion; the most eye-catching one is still Google Cloud - revenue soared 82% to US$24.8 billion, operating profit margin nearly doubled to 35.6%, and backlog of orders currently reaches US$514 billion. Gemini also made progress: monthly active users reached 950 million, and the model API processed up to 22 billion tokens per minute. The cloud business can meet massive demand while expanding profit margins. This is the direct reason why Google dares to double capital expenditures. This torrent of AI infrastructure extends far beyond Google. Morgan Stanley predicts that capital expenditures by hyperscale cloud service providers will jump from approximately US$800 billion in 2026 to US$1.2 trillion in 2027; Goldman Sachs Research estimates that it will be US$765 billion in 2026 and reach US$1.6 trillion in 2031. McKinsey's caliber is even more exaggerated - the total related capital expenditures by 2030 will be 6.7 trillion US dollars, of which AI-oriented data centers alone will swallow 5.2 trillion US dollars. When a company's quarterly capex can double year-on-year and cloud profit margins double simultaneously, it is obviously not betting on returns in one or two quarters, but on an entire infrastructure cycle rewritten by AI. via AI News (author: AI Base)