Nvidia disclosed Wednesday that its purchase commitments to suppliers for AI chip components reached $279 billion last quarter, more than doubling from $119 billion. The commitments, primarily for high-bandwidth memory, aim to secure supply for its advanced AI chips and support a forecast of 70% revenue growth for the fiscal year ending January 2028. Chief Executive Jensen Huang said supply constraints remain the only barrier to faster growth. Chief Financial Officer Colette Kress warned gross margins will dip to about 71.5% next quarter due to higher memory prices.

The strategy carries historical risk. Cisco Systems took a $2.2 billion inventory charge in 2001 when the dot-com bubble burst and supplier commitments became liabilities. Nvidia is also extending financial backstops to customers, including a $105 billion guarantee on an OpenAI data center lease, up to $125 billion in residual-value support within a $500 billion financing deal, and $36 billion in sales guarantees for cloud-computing firms. It also disclosed $20 billion in data center leases intended for transfer and "selective credit enhancement" for nearly two gigawatts of computing power for an unnamed frontier AI lab.

Analysts note a downturn could squeeze Nvidia from both suppliers demanding payment and customers defaulting on guaranteed deals, while equity investments in its own customers could decline in value.