Nvidia faces scrutiny over its artificial intelligence deals

Nvidia faces fresh scrutiny over the scale and structure of its artificial intelligence business, after commitments tied to AI passed 750 billion dollars and intensified concerns about financing circular, a model critics say blurs the line between investment and customer support. The issue has sharpened as the company announced a string of deals that connect chip sales, equity stakes and infrastructure funding across the AI supply chain.

In recent days, the company led by Jensen Huang has unveiled or discussed several large-scale commitments. One of them is a partnership with South Korea’s SK Group worth more than 500 billion dollars, combining Nvidia’s memory chip purchases from SK Hynix with SK Group’s acquisition of Nvidia supercomputers. Huang framed the deal as evidence of the massive business opportunity created by the next wave of digital infrastructure.

The pressure increased after reports that Nvidia is negotiating a guarantee of as much as 250 billion dollars for OpenAI’s lease on a huge data center hub in Ohio, while also discussing financing 350 billion dollars in chip purchases for the same project. Nvidia has also confirmed a significant investment in Safe Superintelligence, the research lab co-founded by Ilya Sutskever, and reports say it is the unnamed tenant behind major Texas data center leases linked to Hut 8.

Warnings from Goldman Sachs and investor Michael Burry have resurfaced. Both have argued for months that tightly linked arrangements could distort demand signals and magnify losses if AI revenue does not catch up with spending. For critics, the issue is not only the size of the commitments, but the closed financial loop they create around them.

Markets reacted sharply. Nvidia shares fell about 5% on Monday, their worst session since early June, and the company briefly lost its position as the world’s most valuable company to Apple. At the same time, the cost of insuring Nvidia debt against default posted a record jump, underscoring growing unease around a strategy that depends on keeping AI infrastructure investment at full speed.

Huang has pushed back. In January, he called the financing circular criticism ridiculous while discussing Nvidia’s investment in CoreWeave, arguing that the company’s capital contributions are only a small part of what its partners must ultimately raise. But Bloomberg Intelligence has taken a more cautious view, saying a slowdown in infrastructure deployment, even for a few months, could be a major problem for Nvidia.