The AI boom is leaning more heavily on foreign capital and circular investment deals, and the financial dynamics are starting to look uncomfortably like the conditions that preceded the 2008 financial crisis. That comparison isn't coming from a single source — it's circulating among market watchers as the scale of cross-border funding and the tight web of deals between the same set of players become harder to ignore.
What 'circular' means in the AI trade
Circular investment isn't a formal term, but it describes a pattern that's easy to spot once you look for it: Company A backs Company B, which turns around and buys services or hardware from Company A, and the money effectively goes in a loop. In the AI space, that loop often runs through cloud credits, chip purchases, and strategic stakes. The result is that reported revenue and growth can be inflated by deals that wouldn't exist without the same small group of funders on both sides of the table.
This isn't fraud. It's a structure. But it means the health of any single AI firm is tied to the health of its investors, its customers, and its partners in ways that are hard to untangle from the outside.
Foreign capital is doing more of the heavy lifting
The AI buildout — data centers, chips, talent — is expensive, and a growing share of the bill is being covered by money from outside the United States. Sovereign wealth funds, foreign conglomerates, and overseas institutional investors have become core backers of AI infrastructure and startups. That's not inherently a problem. Foreign capital has funded American technology for decades.
The concern is what happens if it stops. If a foreign investor pulls back because of domestic economic pressure, a currency crunch, or a change in political priorities, the AI companies relying on that money don't have an easy substitute. The same is true if multiple foreign backers retreat at once, which is exactly the kind of correlated behavior that turned subprime mortgages into a global crisis in 2008.
Why the 2008 comparison keeps coming up
The parallel isn't about subprime mortgages or credit default swaps. It's about opacity and interdependence. Before 2008, a lot of risk was hidden inside complex instruments that few people fully understood, and the exposure was concentrated among a handful of institutions that all assumed the others were safe.
Today, the AI trade has a similar feel: a small number of very large players, a lot of money moving between them, and a widespread assumption that demand will keep growing. If that assumption breaks — if AI adoption slows, if regulation bites, if a major funder steps back — the unwind could be fast and messy, because the same circular deals that look like strength during the boom can become a channel for losses during a downturn.
There's no single regulator or rule that governs these cross-border, cross-company relationships. That's part of what makes the comparison to 2008 resonate. The risk isn't in one place. It's in the connections.
What to watch from here
No one is calling this a crisis. The AI boom is still generating real revenue and real products, and the companies at the center of it are among the most valuable in the world. But the funding structure deserves scrutiny. Watch for disclosures about who is backing which AI firms, how much of the money is coming from outside the US, and whether the same investors keep showing up on both sides of deals.
The next concrete signal may come from quarterly filings or from foreign investment authorities in countries whose funds are heavily exposed. Until then, the comparison to 2008 will stay in the conversation — not as a prediction, but as a warning about what happens when you can't see where the risk actually sits.
