AI Infrastructure Spending Set to Hit $1.13T by 2027
Wall Street's estimate for top-5 hyperscaler capex in 2027 has more than doubled in 12 months — from $495B to $1.13 trillion. China's AI-related export surge adds a geopolitical dimension to the buildout.
What we know
An account on X RTs the staggering revision: just 12 months ago the 5 largest hyperscalers were expected to spend ~$495B in 2027; today that estimate is $1.13 trillion. This directly connects to the FT's reporting that China's AI-related shipments are helping drive the export surge. On the corporate response side, Xbox Labs was announced as Microsoft's answer to AI-driven game remixing — praised across tech/crypto Twitter as unusually forward-thinking. Scale AI CEO Alexandr Wang is casually plugging AI creative tools. On the dev side, another account on X describes nanocode(x) as a new paradigm in rust/wasm harness aggregation.
How the story developed
Wall Street's estimate for top-5 hyperscaler capex in 2027 has more than doubled in 12 months — from $495B to $1.13 trillion. China's AI-related export surge adds a geopolitical dimension to the buildout.
Wall Street voices are raising serious doubts about AI economics: Jefferies draws a 2008 subprime parallel, Carlyle says the AI buildout is worth only ~15% of its $2.5T backlog, and Anthropic's revenue run rate reportedly peaked in July 2026.
Sources
“China's exports surge 36% as AI-related shipments boom”
View post →By the Global Eye 24 desk, with AI assistance. Updated as the story develops. Spot an error? Write to corrections@globaleye24.com.