NVIDIA is working to organize a massive financing arrangement with some of Wall Street's largest firms to help fund the ongoing buildout of AI infrastructure. According to reporting from Bloomberg and the Financial Times, the effort involves Apollo Global, BlackRock (through its Global Infrastructure Partners unit), Blackstone, Brookfield Asset Management, Goldman Sachs, and KKR, and is aimed at mobilizing over $500 billion in third-party capital over time. Sources differ slightly on how far along the deal is: one report describes signed memorandums of understanding establishing a compute financing platform, while another characterizes it as NVIDIA still working with the firms to put the package together.
The financing would reportedly allow hyperscalers and neocloud companies to secure debt-based financing for NVIDIA GPUs and data center hardware, with the equipment itself potentially serving as collateral. This approach could help NVIDIA avoid taking on additional risk itself through circular financing arrangements with partners such as CoreWeave. Despite the scale of the announcement, NVIDIA's stock fell more than 3% after the Financial Times first reported the story.
The financing push comes as capital needs across the AI sector continue to climb. NVIDIA issued $25 billion in bonds in June, its first bond sale since 2021. Big tech companies are on track to spend more than $730 billion combined on AI this year. NVIDIA is also reportedly in talks to provide roughly $250 billion in financing guarantees for OpenAI's planned data center campus in southern Ohio, as well as up to $350 billion to help fund OpenAI's chip purchases. Separately, Intel announced a $15 billion stock offering to fund its foundry expansion, and TSMC reported July revenue of roughly $14.49 billion, up 5.6% from June and 44.7% year-over-year, reflecting rising customer commitments rather than new capacity coming online.
At the same time, China is tightening control over its AI talent pool. According to Bloomberg, China plans to impose travel restrictions on AI engineers at strategically important companies such as Alibaba and DeepSeek starting September 15, requiring government authorization for any foreign travel. The move may be linked to a separate development in the U.S., where the FCC is reportedly considering a ban on Chinese-made optical transceivers used in fiber-optic data transmission.
China is also expanding its AI infrastructure domestically, building large data centers in the Inner Mongolia desert, where abundant solar energy helps lower operating costs. Together, these developments illustrate an accelerating divide between the U.S. and China in AI development, with each side moving to protect its own strategic advantages—the U.S. through massive private financing for infrastructure, and China through tighter control of talent and new domestic data center capacity.