NVIDIA transforms AI factories into new class of assets with support from Apollo, BlackRock and KKR

NVIDIA announced a partnership with six of the largest financial institutions in the world — Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR — to enable an independent financing model for AI factories. With this, AI computing ceases to be seen only as operational cost and becomes treated as a class of Investible Assets, paving the way for institutional investors to participate in the expansion of this infrastructure.

The initiative, revealed in the official NVIDIA blog, points to a time when the demand for AI processing exceeds the installed capacity. Instead of relying exclusively on the balance sheets of big techs, the construction of specialized data centers can be financed by private capital, with a return linked to the operation of these plants.

AI factories: what are they and how do they work?

AI factories are new generation data centers, specifically designed to train and execute artificial intelligence models. Unlike traditional clouds, they are built around NVIDIA GPUs, such as Hopper and Blackwell superchips, and optimized for tasks such as inference and training of generative models.

NVIDIA has been calling these facilities "AI factories" to emphasize that computing has become a production input: just as a factory transforms raw material into product, an AI factory transforms data into intelligence.

IA plant investors: Apollo, BlackRock and KKR

The list includes weight names of the global financial market. Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR are managers who together manage trillions of dollars in assets. The proposal is that they act as independent financers of AI factories, creating an ecosystem in which computing infrastructure can be built, operated and remunerated by large technology companies.

This separation is significant. By making AI computing a class of assets, the model allows pension funds, insurers and other long-term investors to allocate capital in a sector with growing demand, without depending on the capital structure of the cloud giants.

Impact of AI plants on developers and businesses

For companies and developers, novelty can mean broader access to processing capacity. With more capital available to build AI plants, the supply of GPUs in the market tends to increase — which, in theory, can push down the prices of computational capacity rental.

In Brazil, where the AI infrastructure is still limited, the trend is that local companies can enjoy global solutions without having to invest in their own data centers. This is a relevant change for the national AI ecosystem, which today depends mostly on foreign clouds.

The movement should also attract the attention of developers who work with language models and generative AI tools. In a scenario of shortage of GPUs, independent funding reduces the risk that innovation will be concentrated only in companies with higher investment power.

Meanwhile, other news shows that the AI sector remains heated: Blacksmith raises $45 million to accelerate code validation generated by AI — a sign that capital is migrating to different layers of the AI chain.

Frequently Asked Questions about AI factories

What are NVIDIA AI factories?

They are data centers specialized in AI computing, equipped with NVIDIA GPUs and optimized to train and execute artificial intelligence models.

How does the financing of AI plants work?

NVIDIA is forming partnerships with institutional investors such as Apollo, BlackRock and KKR to create an independent financing model, in which the infrastructure is built and operated by third parties with private capital.

Does that reduce the cost of AI computing?

Potentially, yes. With more capital to build AI factories, the processing offer tends to grow, which can increase the offer and reduce GPU rental prices in the long run.