What happened
Over the next five years, US data center capacity is projected to grow from 25 to 70 gigawatts, part of a global buildout costing roughly $5t. This buildout will require about $4t in new debt, which represents a 34% expansion of the US corporate bond market.
Why it matters
The scale of this debt is unprecedented, tripling the outstanding commercial paper market and exceeding the global private credit market. It also equals 91% of the US municipal bond market, raising the question of whether municipalities will use such bonds to fund data centers, as they did for power plants.
What to watch
The test is whether AI revenue growth can close the gap from $100b to $200b today to $1.2t–$1.5t by 2030, since the $4t debt buildout hinges on servicing that obligation. Watch the 55% CAGR required versus today’s 37%–82% hyperscaler growth.
Summaries like this, in your inbox every morning.
The financing of AI data centers is shifting from a venture capital story to a macroeconomic credit event. The required $4t in new debt is substantial, rivaling the size of major credit markets. The comparison to the municipal bond market is telling—it suggests that municipalities might play a role in funding these projects, as they have historically for infrastructure like power plants. The revenue projections highlight the economic stakes: achieving the necessary growth rate will depend on sustained demand for AI services, and the current acceleration among hyperscalers like AWS, Azure, and Google Cloud provides some confidence, though the gap between current and required growth rates remains significant. This buildout could reshape credit markets and local government finance, making it a critical issue for business readers to monitor.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
OpenAI's models, including GPT-6 Astra, and its Codex coding agent will be available through Google Cloud's Mo…
Anthropic reportedly plans to spend $750 billion on adding compute capacity in the coming years, including hun…
NVIDIA announced DSX POD, a modular AI factory pairing Vera Rubin systems with power, cooling and networking…
On held-out synthetic data (32 requests, 7 tasks), sending everything to an LLM cost about $0.15, versus $0.01…
A survey conducted by Google and other parties found that 58% of respondents in Southeast Asia use generative…
OpenAI's venture arm and Peter Thiel's Founders Fund invested in Coherence Neuro, a startup developing a brain…
