Nvidia’s $500 billion AI infrastructure push leaves crypto compute further behind

Nasdaq-listed chipmaker Nvidia (NVDA), the bellwether for everything AI, is pushing Wall Street banks to treat its AI computing power like commercial real estate, toll roads or power plants: as an investable infrastructure asset.

Nvidia said Monday it has signed memorandums of understanding with six Wall Street heavyweights -- Apollo Global Management, Blackstone, BlackRock, Brookfield Asset Management, Goldman Sachs and KKR -- to set up financing platforms that could eventually tap more than $500 billion in third‑party capital.

The goal, according to the chipmaker, is to treat AI compute as a bankable infrastructure asset rather than a pure tech expense, encouraging customers to build out AI data centres and lock in demand for Nvidia's hardware.

"This is really the first time that technology chips have become an investable asset class. These are revenue-generating assets now. They're productive, they're long-lived, they're fungible, they're flexible," Jensen Huang, NVIDIA's founder and CEO, said.

"Fundamentally, what's different about this industry and this way of doing computing is that the computer is now part of the infrastructure, like electricity, like the internet, and so you have to think about it like it's infrastructure," he added.

What's AI compute

AI compute refers to the raw processing power used to train and run artificial intelligence models. Specialized chips, mostly Nvidia's high-end GPUs, primarily do that work and make up the large data centers that Nvidia calls "AI factories."

These factories consume electricity and data to generate useful intelligent systems, such as those that power chatbots, generate images and videos, help design new drugs, drive robots, and run countless other applications. The more advanced the AI, the more of this specialized computing power it needs.

The change that Nvidia seeks

As of today, most companies view buying or renting computing power as a straightforward technology expense that sits on the balance sheet and loses value quickly as newer chips arrive. (Imagine your iphone getting outdated with newer versions arriving in the market)

But, according to Nvidia, this view is outdated, because its systems are widely adopted and can be used by several customers, generating an income stream over years. Hence, the AI factory should be treated as a long-term investable asset.

Here's an example of the change Nvidia is pushing for.

Let's say a company needs powerful AI chips right now. It will most likely spend millions of dollars of its own cash or take out a regular business loan from a bank to buy a large cluster of Nvidia GPUs. This expense now goes on the balance sheet as equipment, which accountants will depreciate over a few years as newer models arrive. The bet is that before these chips lose most of their value, the AI product built with them will generate enough revenue to more than offset the cost.

Now, the change is that the same company still needs the chips but is no longer shelling out the full amount out of its own pocket. That is done by a financing platform backed by big institutional investors. In return, they look to earn money from the rental income the chips generate over many years.

Essentially, the company has accessed the processing power in AI factories by paying rent. Institutional investors are willing to do this because the same Nvidia chips can serve many different clients and workloads, generating a steady stream of rental income.

This shift is at the heart of Monday's announcement.

Under the MoUs, Wall Street banks will independently assess each project for customer demand, expected utilization and cash flow before deploying capital. In some deals, Nvidia may cover 25% of the risk if the chips lose value, but the main point is that the lenders still do their own checks and decide on each project themselves.

Nvidia's initiative could open up a larger pool of long-term capital at a time when investors have begun to question whether the massive capital spending by big tech firms on AI will deliver returns.

"Every industrial revolution has been built on infrastructure: electricity, transportation, communications and computing, with every buildout enabled by external financing. AI factories are the infrastructure of the intelligence era," Huang said.

Where do decentralized compute networks stand

While Nvidia has become a centralized AI powerhouse, networks such as Akash and Render have looked to create a global marketplace for computing power run by ordinary people and coordinated by blockchain.

They have grown, yet they have not matched Nvidia's scale. Research highlights clear limits.

Epoch AI finds that the largest active decentralized training networks still deliver only about one-three-hundredth the throughput of frontier data centres, noting that "it's unlikely that decentralized developers will amass frontier amounts of compute this decade."

Low internet bandwidth forces GPUs to spend most of their time waiting for data rather than computing, while the need for cryptographic verification of every result imposes a heavy extra cost.

Industry analyses also flag the absence of corporate-grade service-level agreements and the practical difficulty of moving massive datasets to scattered machines as major barriers to enterprise adoption.

These physical, economic and operational constraints explain why decentralized networks lag the likes of Nvidia and the gap is set to widen with the chip maker's MoU with Wall Street banks.

CoinDesk reached out to top decentralized compute networks like Render and Akash for a comment on the matter.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.