Three headlines from the same week tell a connected story. Nvidia is investing $1 billion in AI coding startup Poolside on top of a $6 billion licensing deal. Broadcom is negotiating as much as $100 billion in debt to build AI infrastructure tied partly to Anthropic. And Anthropic itself is reportedly preparing an IPO built on annualized revenue that depends heavily on continued enterprise AI spending. Pull on any one thread and the others move with it.
This is the dynamic critics call circular financing: chipmakers and cloud providers extending capital, credit or investment to the very companies that buy their products or license their technology, then pointing to the resulting revenue as evidence of organic demand. Supporters, including Nvidia's own leadership, argue this framing misses the point — that the underlying compute demand is real and contracted, and that vendor financing has always played a role in capital-intensive industries building out new infrastructure.
Skeptics counter that heavy leverage across interconnected companies can mask how much of the AI boom's revenue growth is being pulled forward or subsidized rather than organically generated, a concern that matters more the closer any of these companies get to public markets. If AI spending growth decelerates even modestly, the same interlocking deals that look efficient today could transmit losses across the ecosystem faster than a less concentrated financing structure would.
Neither view is obviously wrong. What's clear is that the AI industry's biggest players are now financially intertwined in ways that make it harder to evaluate any single company's growth in isolation.
