Nvidia's $500B 'compute as asset class' plan draws mortgage-backed security fears

Nvidia is spearheading a sweeping financial strategy to position compute as a new investable asset class. CEO Jensen Huang announced partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to assemble $500 billion in financing. "This is really the first time that technology chips have become an investable asset class," Huang told CNBC, calling the chips "revenue-generating assets" that are "long-lived, fungible, and flexible." BlackRock CEO Larry Fink compared the move to the early days of mortgage-backed securities in the 1970s, hinting at a new era of financial engineering.

Yet the article highlights glaring contradictions. Just last year, Huang said of Nvidia's previous-generation Hopper chips: "When Blackwell starts shipping in volume, you couldn't give Hoppers away." That undercuts the narrative that chips are durable, appreciating assets. Skeptics also point to signs of AI overcapacity: data center saturation, Chinese open-source models requiring far less compute, and unresolved questions about whether frontier labs like Anthropic and OpenAI can turn a profit. The $500B plan is also only a memorandum of understanding—not a binding deal. Notably, Nvidia's earlier $100B MoU to invest in OpenAI quietly fell apart. The article suggests this is more trial balloon than firm commitment.