Nvidia Corp. is set to lease a major data centre in Nevada backed by a $3.8 billion junk bond sale, marking the latest sign of the borrowing surge tied to artificial intelligence infrastructure.
An entity backed by alternative asset manager Tract Capital plans to issue the high-yield debt to finance part of the construction of a 200-megawatt data centre and substation in Storey County, Nevada, according to people familiar with the matter. The bond offering was reportedly increased by $150 million amid strong investor demand, with pricing discussions pointing to a yield of around 6%.
The planned facility highlights how developers are increasingly tapping the high-yield market to fund AI-driven projects. While investment-grade borrowers such as Oracle Corp. and Meta Platforms Inc. have raised billions to expand data capacity, more speculative issuers are now following suit to capitalise on AI momentum.
Nvidia will start its first lease under the new agreement which will continue for 16 years and the company can choose to extend the lease for two more periods of 10 years each. The company which develops advanced AI chips and systems needs to expand its computing resources because model development and chip design work have increased in complexity and require more power.
Tract Capital which manages approximately 6 billion dollars in assets that focus on digital infrastructure has introduced its Fleet strategy which aims to construct new facilities dedicated to hyperscale customers.
The Fleet I fund serves as the first fund of the company which owns the borrowing entity that issued the bond sale. The fund will provide around 620 million dollars in equity to the project but additional debt proceeds will decrease that equity share.
JPMorgan Chase & Co. which leads the transaction together with Morgan Stanley and other banks participate in the deal which will reach its pricing stage soon.
Nvidia runs its own data centers while also renting extra data center space from major cloud services providers including Amazon Web Services and Microsoft Azure because AI training and deployment need continues to grow.