Nvidia says AI computing is a financial asset and raises half a trillion dollars to prove it

Nvidia announced partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR aimed at raising over $500 billion of third-party capital for building AI infrastructure over time. The move shifts the industry from project-by-project chip purchases to a model where AI factories are financed as production infrastructure with recurring financing platforms and a diverse customer base that uses compute to generate revenue.

The core claim from Nvidia is that its compute does not lose value like ordinary hardware but improves over time thanks to its CUDA software. As an example, the A100 – a chip from the Ampere family launched in 2020 – remains in active commercial use today for training, fine-tuning, inference and high-performance computing. According to Nvidia, customers continue to commit to long-duration purchase periods, extending the economic life of the A100 to close to a decade. This is the manufacturer’s claim; it is accurate that the A100 is widely used, but it omits the question of whether the price per GPU hour stays stable as demand moves to newer architectures.

Nvidia’s pricing data shows an increase in H100 rental rates per year, from $1.70 per GPU-hour in October 2025 to $2.35 in March 2026. The median on-demand price across providers rose from $2.00 per GPU-hour in October 2025 to $2.70 in June 2026. Blackwell capacity trades at a premium, with B200 cloud prices ranging between $5.30 and $7.05 per GPU-hour. These figures are meant to demonstrate the resilience of the economics around Nvidia compute, but they reflect provider rates rather than Nvidia’s direct costs and are influenced by supply shortages as well as organic demand.

The financial logic Nvidia builds is that an Nvidia-based AI plant is an income-generating asset, serving a broad market, improving performance over time, and able to be redeployed to another customer or cloud provider. The standard architecture used by the major clouds is expected to create a deep market of potential users and end-users (offtakers), thereby protecting residual value. Nvidia concludes that this constitutes a type of infrastructure asset suitable for long-term institutional investment.

Demand for AI infrastructure exists, but access to capital is not evenly distributed. Nvidia notes that many AI firms, organizations and AI clouds need compute but cannot obtain financing at the scale or cost required for rapid build-out. Partnerships with institutional capital firms are intended to bridge this gap through recurring financing platforms that help the AI system build the factories it requires.