Blackstone’s Jas Khaira says fears over AI-buildout debt are overblown

Your AI infrastructure costs could increase as lenders charge more for compute loans.

Blackstone's Jas Khaira stated that fears regarding debt accumulated for AI infrastructure buildouts are exaggerated. The private equity firm is heavily involved in the AI sector, offering loans to data centers and investing in AI models through its dedicated hub, Blackstone N1 (BXN1). Khaira highlighted Blackstone's significant ownership of compute infrastructure, positioning them as a major player beyond hyperscalers. He explained their financing model for neoclouds, like CoreWeave and Firmus, which leverages customer contracts and GPUs as collateral. Despite recent market jitters, including CoreWeave's stock drop and Firmus's canceled IPO, Khaira asserts that the underlying strength of the customers, such as Microsoft and Meta, mitigates risk. He also pointed to the ongoing demand for compute as a fundamental driver of data center construction.

Khaira's assertion that fears are overblown suggests that despite recent market volatility and the cancellation of Firmus's IPO, Blackstone continues to see value and manageable risk in AI infrastructure debt. This perspective implies that large-scale AI projects will continue to receive substantial, albeit more expensive, financing from major financial institutions. The firm's strategy of leveraging creditworthy customer contracts as a primary risk mitigator for loans, even to companies with limited profitability, indicates a confidence in the sticky nature of enterprise AI adoption.

This approach signals a market where compute capacity remains a critical bottleneck, driving continued investment in data centers and GPU providers. However, the rising cost of capital, evidenced by wider credit spreads, means that the economics of AI buildouts are becoming more challenging. Companies reliant on this debt will likely face higher operational expenses, which could eventually trickle down to end-users or impact the pace of new AI service development.