The traditional “next big asset” stock market investing strategy was to invest in the products that have been around for awhile. Gold, real estate, and, naturally, crypto.

But there's something out of the ordinary going on in the background.

In fact, the devices that drive AI — and more specifically powerful GPUs — are beginning to feel like a valuable commodity in itself.

More than just a piece of technology.

Not a Gamer's AddOn for their PC!

A useful resource that provides an income.

And that might subtly shift investor logic on investing in the AI boom.

GPUs Are Becoming the Engines of the AI Economy

Artificial intelligence needs an enormous amount of computing power.

Large AI models, inference, image/video generation, data processing, and AI agent operations are demanding and require significant hardware infrastructure.

GPUs are at the heart of most of this demand.

It's interesting to note that it is not always necessary for a company to own all of the GPUs it uses in order to have demand for computing power. As with offices, companies can now increasingly lease access to GPU computing power when they need it.

This is a potential strong economic model.

You can buy a GPU, install it in the data center, and then lease it to a developer, startup, researcher, or company developing an AI product.

In simple terms:

Computing power requires hardware and increases demand for renting hardware, which in turn increases revenue.

That's more a purchase of infrastructure than a purchase of a gadget.

From Gaming Hardware to Productive Capital

A GPU's worth had traditionally been limited to what one individual could achieve with it.

Play games.

Edit videos.

Render 3D graphics.

Harvest cryptocurrencies, when the market is favorable.

However, AI has brought another big use case to the fore: GPUs can now offer a service for computing.

This makes interesting comparison with other productive assets.

A house can make money at the rental business.

Products can be made on a factory machine.

Websites can be installed on a server.

And a powerful GPU can potentially be used to provide computing resources.

There is a huge difference though – technology’s expiration date is vastly different from real estate’s.

A building could be in use for decades. It may be a lot sooner than that before a newer and more efficient piece of hardware begins to compete with a GPU.

So, GPUs also are not a one-to-one replacement of traditional assets.

However they could be a new class of digital infrastructure assets.

The AI Boom Is Creating a Compute Shortage

The most pressing issue for AI firms is having the required amount of computing resources.

The cost of building AI models is high and as more companies enter into the AI race, the demand for high performing chips has accelerated.

The economics come into play here.

Access becomes valuable when demand for computing power grows faster than the infrastructure to support it.

It's also been observed in cryptocurrencies.

In periods of bull markets, there was a huge demand for mining machines and computation energy. The value of hardware was created as it became a direct part of an expanding digital economy.

This new trend may be extended with the help of AI.

GPUs can be used for thousands of AI applications, rather than computing power being used to secure a blockchain or to mine digital assets.

Chatbots.

Image generators.

Video models.

Autonomous agents.

Scientific research.

Robotics.

Financial analysis.

And applications that likely haven't even been thought of yet.

Could GPU Ownership Become More Accessible?

This is where it may get quite interesting for the crypto and blockchain community.

The concept of tokenization of real-world assets and infrastructure has been tried on blockchain technology.

Suppose the entire pool of computing infrastructure is digitally represented, and the investors can access the revenue generated by AI infrastructure.Suppose that large pools of computing hardware are digitally represented, and thus the investors can access the revenue generated by AI infrastructure.

A person might potentially get exposure to a part of a disproportionately big computing network without having to purchase and run an expensive GPU.

It is too early and there are clear risks with these models.

However, the concept seamlessly integrates with the larger vision of real-world assets going on-chain.

There have already been increased interest in tokenized treasuries, real estate, private credit and other financial assets.

May be computing infrastructure one day be added to the list?

It's no longer an impossible feat.

The Biggest Risk: Hardware Doesn't Stay Valuable Forever

While we're still in the process of declaring GPUs "digital gold," there is a major caveat.

A new version of gold doesn't mean that the old version is no longer valid.

GPUs do.

A high-performance chip can become a big loser if future generations of chips can perform better or at lower energy use.

There are other risks, too:

Electricity and cooling expenses
The price of renting GPUs is decreasing.
Rapid hardware depreciation
Excess capacity of computing infrastructure.
Competition from specialized AI chips
The reliance on a handful of large technology firms.The high degree of reliance on a few large tech companies.
Changes in AI demand

This is not just simply purchasing Bitcoin and storing them in a wallet.

Maintenance, power, cooling, customers and constant upgrades are necessary for GPUs.

They are productive assets, but productive assets need management!

The Bigger Story Isn't the GPU Itself

The coolest aspect of this development could be graphics cards themselves.

It could be because computing power is becoming a market place.

In a future where compute is a commodity as widely traded as cloud storage or bandwidth, it will be very easy to purchase and sell.

Need AI processing power for 2 hours?

Rent it.

Have unused infrastructure?

Lease it out.

Looking for AI infrastructure without a data center?

There are new financial models that in the future may allow that.

This may generate a whole new strand of the digital economy.

Ownership, payments, coordination, and use of infrastructure could be enabled by crypto in the future.

Final Thoughts

The concept of scarcity in the digital world was first conceived on Bitcoin.

The concept of programmable finance was introduced to the financial world by DeFi.

But now, AI is driving huge demand for something more tangible: computing power.

This might render GPUs and other AI infrastructure more important as an economic asset.

Unlike traditional assets, however, their value is very much technology, demand, energy expense and the rate of change in the industry.

GPUs might not be the gold of the future, then.

This analogy is more apt than that:

They are emerging as the epitome of the AI economy's factories.

Like all great industrial waves that have come before it, the AI revolution could be generating a need for new infrastructure.

This is different, however, because this time, it's in the form of chips, compute and algorithms.

But the scramble to own it has already started.