The artificial-intelligence investment story may be entering a new phase.

For much of the AI boom, investors have focused heavily on the companies supplying the essential computing infrastructure. Nvidia became the clearest symbol of that first wave, while memory manufacturers such as Micron benefited from the enormous demand generated by AI data centers and increasingly sophisticated computing systems.

But the next group of winners may look very different.

A new investment thesis highlighted by Yahoo Finance argues that the AI opportunity is expanding beyond the companies that manufacture the chips and memory used to train and operate advanced models. As AI moves from experimentation into widespread commercial deployment, investors are increasingly looking for businesses that can benefit from the technology's broader economic impact.

That represents an important change in the way the AI trade is being viewed.

The first phase of the AI investment cycle was largely about infrastructure. The rapid development of generative AI created enormous demand for advanced processors, high-bandwidth memory, networking equipment and data-center capacity.

Nvidia became the market's most prominent beneficiary because its accelerators became central to the training and deployment of large AI models. Memory manufacturers also benefited as AI systems required increasing quantities of high-performance memory.

But infrastructure investment cannot expand indefinitely at the same pace.

As companies build more computing capacity, the focus inevitably begins shifting toward what that capacity can actually accomplish. Businesses capable of using AI to increase productivity, automate processes, improve customer service or create entirely new products may become the next beneficiaries.

That creates a much broader investment universe.

Instead of asking which company manufactures the fastest processor, investors can begin asking which businesses can use AI to reduce costs or increase revenue.

The transition could have significant implications for sectors outside traditional technology.

Financial companies can use AI to analyze enormous amounts of data and automate routine operations. Healthcare businesses can apply machine learning to research, diagnostics and administrative tasks. Industrial companies can use AI for predictive maintenance, robotics and supply-chain optimization.

Retailers can use the technology to personalize recommendations and improve inventory management. Software companies can integrate AI agents directly into business workflows.

In each case, the potential value comes not necessarily from selling AI hardware, but from using AI to make an existing business more efficient or more valuable.

This is why the next AI winners could look substantially different from Nvidia and Micron.

The investment opportunity may increasingly depend on the economics of AI adoption rather than simply the economics of AI infrastructure.

One of the biggest questions for markets is whether AI investment eventually produces measurable productivity gains across the economy.

Companies have already committed enormous sums to computing infrastructure, but investors ultimately need to see a return on those investments. If AI allows companies to accomplish more with fewer resources, margins could improve and earnings could rise.

That would potentially create a second stage of the AI trade.

In the first stage, infrastructure suppliers capture the spending associated with building the AI ecosystem. In the next stage, companies that successfully deploy that infrastructure could capture the economic benefits.

The distinction is important because it could dramatically broaden the number of companies participating in the AI boom.

The expansion of the AI opportunity does not mean every company adding an AI label to its strategy will become a winner.

Investors still need to determine whether AI produces meaningful financial benefits. A company can announce an AI initiative without generating additional revenue or improving profitability.

The strongest beneficiaries are likely to be businesses where AI can be integrated into existing operations in a way that creates measurable economic value.

That could make the next stage of the AI investment cycle more fundamental than the first.

Rather than simply rewarding companies associated with the AI theme, investors may increasingly reward companies that demonstrate actual improvements in productivity, revenue growth and margins.

The result could be a market in which the biggest AI winners are no longer concentrated exclusively among semiconductor and memory manufacturers.

Nvidia and Micron helped define the infrastructure phase of the AI revolution. The next phase could be about everything built on top of that infrastructure.

For investors, that means the most interesting AI opportunities may increasingly be found outside the obvious names — in businesses that can turn artificial intelligence from an expensive technology project into a genuine source of competitive advantage and earnings growth.