The next stage of the AI race may depend on which companies can secure long-term capital at lower cost and greater scale. According to Sina Finance, Nvidia joined Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR in August to establish an AI infrastructure financing platform intended to mobilize more than $500 billion.

Analysts cited in the article estimate that capital expenditure by the five major U.S. hyperscalers could reach $730 billion in 2026, with up to one-third expected to be financed through debt. Nvidia’s platform aims to bring outside investors, including pension funds, insurers and sovereign wealth funds, into AI infrastructure financing. The article notes that GPU technology can become obsolete faster than traditional infrastructure assets, making future returns and cash flows less predictable.

The euro area has about €10 trillion in household bank deposits, but European institutional investors tend to favor U.S. technology company bonds over financing local AI startups, the article says. Analysts argue that access to capital, financing costs and risk-sharing mechanisms could increasingly shape AI competition alongside model development and chip performance.