Goldman Sachs takes the lead in bridging institutional capital Nvidia (NVDA.US)'s $500 billion AI infrastructure financing: from “selling chips” to “selling assets”

The Zhitong Finance App learned that in the context of the generative AI wave sweeping the world, computing power infrastructure is transforming from a “cost center” for technology companies to one of the most attractive “asset classes”. Nvidia (NVDA.US), with its GPU at its core, is trying to leverage an unprecedented capital movement — the $500 billion AI infrastructure financing plan is accelerating, and Goldman Sachs, the top Wall Street investment bank, is playing the role of a key “bridge builder.”

On August 10, Nvidia announced the establishment of strategic partnerships with top global organizations such as Apollo Global Management (Apollo Global Management), BlackRock (BlackRock), Blackstone (Blackstone), Brookfield Asset Management (Brookfield Asset Management), Goldman Sachs, and KKR to jointly establish an independent financing platform. The plan's core goals are clear: to turn Nvidia-enabled AI infrastructure into a new, investable asset class, and gradually shift financiers from tech companies themselves to a wider range of institutional investors.

According to previously disclosed information, Nvidia may provide financial support of up to 125 billion US dollars for potential transactions, accounting for about 25% of the total scale, and use this as a “safety cushion” to attract more third-party capital to enter the market.

Goldman Sachs: Full chain service from “inferior capital” to “open bond market”

According to people familiar with the matter, Goldman Sachs is actively discussing participation in this financing plan with potential investors. As one of the six founding partners of the program, Goldman Sachs's role goes far beyond referring clients.

According to the report, Goldman Sachs can provide inferior capital and private credit financing through its asset management business division, while its investment banking department will assist in allocating debt instruments to private credit funds and ultimately connect with the open debt market. This means that Goldman Sachs is building a complete financing chain from private placement to public offering, from equity to claims, and from inferior to priority.

In terms of investor structure, US insurers, asset managers, and banks are expected to form the core investor base of the program, while asset managers are expected to hold a significant share. Goldman Sachs has widely communicated with various investors such as banks, asset management companies, insurance companies, and private credit institutions about this type of structure.

Nvidia's “asset-light” transformation and ecological moat

The strategic importance of this financing plan should not be underestimated. For Nvidia, introducing third-party capital to build AI infrastructure can not only ease the pressure on its own capital expenditure, but also further consolidate its CUDA ecosystem and GPU dominance in the computing power market — whoever invests in the construction of a data center with Nvidia chips as the core is more likely to bind to Nvidia's technical route for a long time.

For institutional investors, AI infrastructure is being viewed as another “supercycle” core asset after the Internet and mobile Internet. Hardware assets such as data centers, computing power clusters, and high-speed internet networks have stable cash flow attributes, complement the high volatility of technology stocks, and meet the allocation needs of long-term capital such as insurance funds and pensions.

The financing scale of 500 billion US dollars is unprecedented in the field of infrastructure investment. If the plan progresses smoothly, it will greatly accelerate the pace of global AI computing power deployment and reshape the ownership structure of data center investments. But challenges also exist: the return on investment cycle of AI infrastructure, energy consumption constraints, technology iteration risks, and concerns about potential excess computing power can all influence institutional investors' final decisions.