The AI boom has a money problem. Not a lack of it, but where it comes from and where it goes.
According to Bloomberg, AI’s global financing loop is drawing billions of dollars from overseas investors. That flood is reviving uncomfortable comparisons to the run-up to the 2008 financial crisis.
The concern is not that AI is fake. The concern is the plumbing: who is lending to whom, how much debt sits underneath, and how hard it is to see the full picture.
The loop, explained
Chipmakers, cloud providers and AI labs are investing in one another, then spending heavily on each other’s chips, compute and infrastructure.
Over half of the incoming investment value into AI firms from 2021 to 2025 came from other AI firms.
A couple of headline deals show how this works in practice. Microsoft and Nvidia hold a combined $15 billion stake in Anthropic. Separately, Nvidia has a $100 billion commitment linked to OpenAI.
Nvidia sells the chips that AI labs need. When the chip seller is also a major financier of the chip buyers, the line between investment and customer acquisition gets blurry.
The debt is growing faster than the clarity
Morgan Stanley anticipates that total AI-related debt could reach $1.5 trillion by 2028. Earlier in the decade, that figure sat in the hundreds of billions.
Private credit means loans from non-bank lenders such as investment funds, which face lighter disclosure requirements than traditional banks.
Estimates put outstanding private credit debt for AI-related firms between $300 billion and $600 billion by 2030. Another projection has private credit to the tech and AI sector rising by another $800 billion, reaching about $1.25 trillion in the coming years.
The ranges vary widely, which is itself telling. When forecasts for the same market differ by hundreds of billions of dollars, nobody has a precise map of the exposure.
Why 2008 keeps coming up
The 2008 crisis was not simply about bad mortgages. It was about leverage stacked on leverage, hidden inside complex structures that tied institutions to each other in ways few fully understood.
Foreign investors control about 20% of US equities, according to the research findings. That level of foreign ownership feeds into current-account imbalances last seen in 2008.
What this means for investors and the industry
For investors holding AI-heavy portfolios, a useful question is how much of a company’s sales come from customers it has also funded.
With Morgan Stanley’s projection pointing to as much as $1.5 trillion in AI-related debt by 2028, balance sheets will matter more than they did when the sector ran primarily on equity.
Private credit loans are typically harder to trade than public bonds. If conditions tighten, holders may find there are few buyers at the prices they expected, which could amplify any sell-off.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.