Key Takeaways
- Broadcom is arranging more than $50 billion to finance custom AI chips for OpenAI, while SpaceX is seeking around $40 billion for Nvidia GPUs.
- Oracle is exploring chip financing through a leasing structure, potentially keeping some borrowing outside its balance sheet.
- Private credit firms including Apollo, Blackstone and PIMCO are becoming increasingly important to AI infrastructure funding.
- The biggest risk is that AI chips lose commercial value before the loans used to buy them are repaid.
- Traders are watching credit spreads, financing agreements and AI chip demand for signs of growing financial pressure.
Artificial intelligence is becoming as much a financing story as a technology story. Broadcom, Oracle and SpaceX are seeking billions of dollars in funding to secure the computing power needed for the next phase of AI expansion.
According to the Wall Street Journal, Broadcom is arranging more than $50 billion for OpenAI's custom chips, SpaceX is discussing approximately $40 billion in borrowing for Nvidia hardware, and Oracle is negotiating financing for a major data centre chip purchase. The proposed deals underline how rapidly AI spending is moving beyond companies' existing cash resources.
The financing could sustain semiconductor demand for years. However, the loans may remain outstanding long after the chips themselves become less competitive.
Why AI Companies Are Borrowing More
Early AI infrastructure spending was largely funded by the operating cash flows of technology giants such as Microsoft, Alphabet and Meta. As data centre projects expanded, companies increasingly turned to bonds, private credit and equipment leasing.
The latest proposed deals show how that financing is changing:
- Broadcom and OpenAI: More than $50 billion in proposed financing for custom AI accelerators, with Apollo and Blackstone among the potential lenders.
- SpaceX: Approximately $40 billion, comprising $30 billion in bonds and $10 billion in loans, to purchase Nvidia GPUs. Apollo is expected to lead the financing, with PIMCO involved in discussions.
- Oracle: An undisclosed financing package involving Apollo and Goldman Sachs, potentially using a separate vehicle to purchase chips and lease them back to Oracle.
Who Is Funding the AI Boom?
Private credit firms are increasingly becoming the financial backbone of AI infrastructure expansion. Apollo appears across several major transactions, including Broadcom's financing arrangements, Oracle's proposed chip purchases and SpaceX's planned borrowing. Blackstone, Goldman Sachs and PIMCO are also involved in various discussions.
These firms provide or arrange capital that helps technology companies secure hardware without paying the entire cost upfront. In return, lenders receive interest payments, while technology companies gain access to computing capacity needed for future growth.
The arrangement benefits multiple parties. AI developers secure infrastructure, chipmakers receive orders, and lenders gain opportunities to earn returns from financing.
However, risk also spreads beyond the technology sector. Private credit funds frequently manage capital on behalf of pension funds, insurance companies and institutional investors. If AI projects struggle to generate sufficient revenue, financial losses could extend to those investors.
What Are the Risks Behind AI Chip Financing?
Three concerns stand out. First, AI companies face substantial operating expenses beyond chip purchases, including electricity, cooling, maintenance and data centre facilities. These expenses must be covered before projects can produce sustainable returns.
Second, suppliers involved in guaranteeing customer financing could face losses if borrowers default. Broadcom, for example, may face both weaker chip demand and financial obligations under any guarantees it agrees to provide.
Third, interconnected financing arrangements could spread financial stress. Nvidia has invested in SpaceX, while SpaceX is seeking debt to purchase Nvidia chips. These relationships can reinforce demand during expansion but create additional exposure if spending slows. Credit markets are already showing caution. SpaceX's credit default swap spreads have risen to record highs as investors reassess its borrowing plans, even as demand for AI computing remains strong.
What Traders Are Watching Next
Traders are watching whether the proposed financing arrangements translate into sustainable AI infrastructure growth or create greater financial pressure across technology and credit markets. The first development is Broadcom's proposed OpenAI financing deal. Its final size, closing date and guarantee terms will help determine how much financial exposure Broadcom assumes.
The second is the movement in credit default swaps and corporate bond yields. Rising borrowing costs for Broadcom, Oracle or SpaceX could indicate that lenders are demanding greater compensation for AI-related risks.
Third, upcoming semiconductor earnings will provide more evidence of whether infrastructure commitments are supporting revenue growth. Broadcom (AVGO) and Nvidia (NVDA) remain important stocks to monitor, while Oracle (ORCL) and SpaceX (SPCX) offer exposure to the companies undertaking significant AI infrastructure investment.
Finally, rental prices for older GPUs will help reveal whether computing equipment is losing value faster than lenders anticipated. Falling rental income could challenge financing models even if overall AI demand continues growing.
Bottom Line
The AI boom is entering a more debt-dependent phase. Large financing arrangements are helping technology companies expand computing capacity while supporting demand for semiconductor manufacturers. However, borrowing introduces risks that strong chip sales alone cannot resolve. Hardware depreciates, interest payments continue, and AI companies must eventually generate enough revenue to justify their investments.
The next test for AI markets is no longer simply how many chips companies can buy, but whether those chips can earn enough to pay for themselves before they become outdated.
For a deeper look at AI financing, private credit and the risks facing chipmakers, read the full article here.