Nvidia Just Teamed Up With Wall Street to Finance $500 Billion in AI Computing Power. Here's What That Means.
Nvidia has signed agreements to mobilize $500 billion in outside money to build AI computing infrastructure, partnering with Wall Street firms including Apollo Global and Blackstone. The goal is to make Nvidia-powered computing more accessible to AI developers, companies, governments, and cloud providers Reuters CNBC. The August 10, 2026 announcement sets up financing platforms through which large investors can fund computing systems powered by Nvidia's chips, called GPUs.
The company confirmed on its newsroom that the money comes from third-party sources, not from Nvidia's own funds NVIDIA Newsroom. The financial firms involved go beyond Apollo and Blackstone, though the others have not been named. Reuters and CNBC both confirmed the agreement structure on August 10, 2026. These agreements are called memorandums of understanding — they signal an intent to work together, but they are not legally binding contracts.
CNBC reported that CEO Jensen Huang called Nvidia's chips an "investable asset" during the announcement CNBC. That means Huang wants investors to think of these chips the way they might think of an apartment building or a toll road: something you buy, that generates income over time, and that you can eventually sell. Huang spoke about the plan on CNBC's Closing Bell Overtime on August 10, 2026, alongside Wall Street leaders involved in the deal CNBC.
The $500 billion figure builds on earlier deals that show the demand behind this push. On July 27, 2026, CNBC reported that OpenAI was talking with Nvidia about up to $250 billion in funding support for OpenAI's plan to lease a new AI data center CNBC. Reuters separately reported on August 8, 2026, that Nvidia planned to invest up to $3 billion in Lancium, a company building the Stargate data center campus in Texas Reuters.
These moves fit together in layers. Lancium's Stargate campus is the physical layer — the buildings and power supply. The OpenAI talks are about the tenant layer — locking in a customer before construction starts. The $500 billion platform announced August 10, 2026, is the money layer — turning future lease payments into investments that outside investors can buy into.
The broader context here is a shift in who pays for and owns AI computing power. In the past, big cloud companies like Amazon, Microsoft, and Google bought GPUs with their own money and kept them on their books. Huang's framing of chips as an "investable asset" would let outside investors own the computing hardware and lease it to AI companies instead. Apollo and Blackstone, both large investment firms, have the money and expertise to set up these kinds of deals at a big scale.
This approach could let AI computing grow faster than any single cloud company could fund on its own. But the details on how risk is shared are not yet public. If newer, faster chips come out and older ones lose value, it is unclear who takes that loss — and that question will affect how much these investments cost.
Nvidia's newsroom also listed a GTC San Jose 2026 session called "Physical AI for Healthcare Robotics - Simulation-First Design & Accelerated Development," pointing to the company's continued push into new fields built on its computing platform NVIDIA Newsroom.
The announcement does not mean $500 billion is ready to spend. The agreements set up a framework and a target for raising money. Actual funding, deal structuring, and project selection will happen over time, and no timeline has been shared yet.
