Nvidia (also known as Hi-Tech) announced on August 10th that it has signed a memorandum of understanding with six major asset management companies including Apollo Global Management, Blackstone, BlackRock, Brookfield, Goldman Sachs, and KKR. According to Nvidia’s official statement, this plan will receive over $500 billion in third-party capital with the goal of establishing a financing platform for AI computing infrastructure, transforming AI computing power into new investable assets.
Regarding the challenges faced by AI infrastructure implementation, Joseph Bae, Co-President of KKR, who is participating in this financing plan, pointed out, “What we have learned is that the delivery is the most difficult part, not the ambition.” Brookfield’s CEO, Bruce Flatt, also stated that computing power is rapidly becoming a crucial cornerstone of modern infrastructure.
Nvidia CEO Jensen Huang elaborated on the underlying logic of this financing plan in the statement. He mentioned that Nvidia’s hardware has high “substitutability and transferability” among different customers and operators, making computing equipment seen as a long-term infrastructure that can generate stable income, which financial institutions can assess for financing.
In the background of technology giants turning to external capital markets for financing, credit rating agency Moody’s industry report pointed out that unprecedented capital expenditures by Hyperscalers are squeezing corporate free cash flow and could potentially increase debt burdens.
According to industry background analysis by CNBC and other financial media, this $500 billion third-party financing platform aligns well with the practical need of tech companies to avoid excessive consumption of their own balance sheets.
Although the official plan positions AI computing power as a long-term financeable asset, based on historical experience in the technology hardware market, GPUs are typically fast depreciating equipment. Whether current AI equipment can maintain long-term mortgage and appreciation functions similar to real estate remains a focus of observation and skepticism in the financial industry. Further asset pricing data is needed to validate this.