NVIDIA Advances AI Infrastructure with US$500bn in Funding

NVIDIA has secured US$500bn in capital with backing from major financial institutions, including Apollo, BlackRock and Goldman Sachs.

The investment will support the company’s plans to develop AI infrastructure, including new data centres capable of housing, running and cooling large numbers of computer chips.

NVIDIA CEO Jensen Huang says “in AI, compute is revenue”. He says the company wants to bring “the world’s leading long-term capital providers together to independently underwrite AI infrastructure”.

The funding reflects a broader shift in how investors view computing capacity. Rather than treating chips solely as technology products, financial institutions are beginning to assess AI infrastructure as a productive asset capable of generating recurring income.

“NVIDIA’s AI factory platform is really an investable asset, an infrastructure asset,” Jensen says.

“The reason for that is because it’s productive, it’s revenue generating, it is fungible, it’s used by just about every cloud service provider, it runs every AI model.”

Can AI chips retain their value?

The strategy depends heavily on NVIDIA’s GPUs maintaining their value and usefulness over time. This is a critical assumption because GPUs, like other technology products, can lose value as newer and more powerful hardware reaches the market.

Traditional asset-backed lending is based on the idea that a lender can recover and sell an asset if a borrower fails to meet its obligations. With GPUs, however, it is not yet clear how long they will remain sufficiently productive to support that model.

The latest AI chips are essential for training frontier models when they first become available. After several years, though, they may be reassigned to less profitable inference workloads.

Ben Emons, Founder of FedWatch Advisors, warns that NVIDIA’s chips could “depreciate faster than expected”, despite their current market value.

He also identifies China as a potential source of pressure for NVIDIA. Chinese companies are building domestic computing capacity at speed, raising the prospect of increased competition and a market flooded with cheaper silicon.

A sharp fall in hardware prices could undermine the collateral behind hundreds of billions of dollars in private loans. If the equipment lost value more quickly than the debt was repaid, lenders and investors could face significant losses, Ben says.

NVIDIA nevertheless remains the dominant supplier of AI chips in the US, with an estimated 75% market share.

The company’s current financial position remains strong. As hyperscalers compete to secure computing capacity, the cost of renting NVIDIA H100 GPUs increased from US$1.70 per GPU-hour in 2025 to US$2.35 per GPU-hour in 2026.

AI investment continues to rise

Google, Meta, Amazon, Microsoft, SpaceX, Tesla, OpenAI and Anthropic are among the organisations using NVIDIA’s chips and GPUs.

Technology companies have invested more than US$1tn in AI projects over the last three years. That spending is expected to continue as businesses expand their use of AI and require additional computing capacity.

The surge in demand has also transformed NVIDIA’s market valuation, which has increased fivefold over the same three-year period.

Jim Zelter, President of Apollo Global Management, says: “Modern compute has emerged as a scarce, mission-critical asset class”.

He adds that continued profitability among companies such as NVIDIA could “drive significant long-term economic growth and productivity gains” throughout the global economy.

Jensen concludes: “Today, we are helping create a new class of productive, investable infrastructure: AI factories.”

With continued support from banks and other investors, NVIDIA will be better positioned to help fund the infrastructure required for the wider AI expansion.