The artificial-intelligence investment boom is entering a new phase as major investors increasingly shift their attention from how much money technology companies are spending on AI to which businesses will ultimately capture the profits.

For much of the rally, Wall Street concentrated on the companies building the infrastructure required for AI: semiconductor manufacturers, networking businesses, data-center operators and the hyperscale cloud providers financing massive computing expansions.

Now the central question is changing.

Investors want to identify the companies that can convert those investments into durable cash flow and earnings growth.

Recent strong results from Microsoft and Amazon have helped ease concerns that enormous AI infrastructure spending could become a drag on profitability. Cloud demand remains robust, while capacity constraints indicate that customers continue to require additional computing resources.

The shift does not mean investors are abandoning Nvidia, Microsoft, Amazon or other major technology companies.

Instead, institutional investors increasingly view AI as a much larger ecosystem.

Chipmakers remain essential because AI models require enormous computing power. Hyperscale cloud companies remain critical because they provide the infrastructure needed to train and deploy those models.

But the next group of winners could emerge among companies that turn that infrastructure into measurable business improvements.

That could include financial companies using AI to automate paperwork, retailers improving pricing and inventory, healthcare businesses automating administrative work, travel companies improving bookings and logistics firms optimizing operations.

The investment opportunity is becoming less about who owns the most GPUs and more about who can generate the largest economic return from them.

Microsoft and Amazon have recently helped restore confidence in the economics of AI infrastructure.

Their businesses benefit from enormous existing cloud operations, meaning they can spread AI-related investments across large customer bases and multiple revenue streams.

That scale matters.

Companies with strong balance sheets, large existing businesses and diversified sources of revenue can afford to invest heavily in AI while waiting for returns to develop.

Smaller companies with limited cash flow may have a harder time doing the same.

Investors are therefore becoming more sensitive to balance sheets and funding structures.

Companies that depend heavily on debt or expensive financing to build AI infrastructure could be more vulnerable if growth slows or capital becomes more expensive.

The next stage of the AI trade may be defined by a simple financial question: does AI investment ultimately produce more cash than it consumes?

For the largest technology companies, the answer is becoming easier to evaluate.

Investors can compare capital expenditure with operating cash flow and examine whether AI-related services are increasing revenue and margins.

Analysts expect operating cash flow growth at the biggest hyperscalers to strengthen relative to capital expenditure growth by 2027, an important signal for investors concerned about the sustainability of the AI buildout.

That could create an important transition.

During the infrastructure phase, investors rewarded companies for increasing capacity.

During the monetization phase, investors may reward companies that demonstrate high returns on that capacity.

Another potentially important group is companies that do not sell AI technology at all.

These businesses may eventually become major beneficiaries by embedding AI into their existing operations.

Travel is one example. Booking platforms can use machine learning to improve recommendations, automate customer support and optimize pricing.

Banks can automate document processing and customer service.

Insurers can use AI to accelerate underwriting and claims processing.

Retailers can use algorithms to optimize supply chains and personalize promotions.

In each case, AI could improve margins without changing the basic identity of the company.

That is why some strategists increasingly believe the next phase of the AI stock-market rotation could favor adopters rather than just builders.

Investors are also looking for businesses that can collect recurring revenue as the AI ecosystem expands.

These companies can be thought of as infrastructure “toll takers.”

Rather than betting on one particular AI application, they provide essential services or components used throughout the industry.

Cloud providers are the clearest example.

Networking, memory, power-management equipment and specialized software can also play this role.

The attraction is that these companies may benefit from AI growth without needing to predict which individual AI application becomes dominant.

As long as overall AI demand rises, the suppliers of essential infrastructure can capture part of that spending.

The enormous amounts of money flowing into AI also create risks.

Not every AI infrastructure provider will survive.

Companies with weak balance sheets, high debt or narrow customer bases may struggle if spending slows or large customers consolidate their purchasing.

That could eventually produce an industry shakeout.

Analysts increasingly expect consolidation as the market matures, with companies that possess diversified businesses, strong infrastructure and sustainable cash flow better positioned to emerge as long-term winners.

The result is a more sophisticated AI investment market.

The early phase rewarded obvious winners such as Nvidia and other companies directly supplying the hardware required to build AI systems.

The next phase could reward a far wider range of companies.

Some will be infrastructure providers. Others will be cloud platforms. Still others will be businesses in completely unrelated industries that use AI to increase productivity.

That means the AI investment story may become less concentrated over time.

It also means stock selection becomes more important.

Investors must distinguish between companies that merely mention AI in earnings presentations and businesses where AI is actually changing revenue growth, costs or margins.

Ultimately, the market is moving from an infrastructure question to an economic one.

Wall Street already knows that companies are spending huge sums on AI.

The question now is who will make the most money from those investments.

Microsoft and Amazon's strong results have helped reduce concerns about the profitability of the current spending cycle, while persistent demand for cloud capacity suggests the buildout still has significant momentum.

But the next winners could increasingly be found outside the traditional AI giants.

They may be the companies that quietly use artificial intelligence to process more transactions, reduce labor-intensive work, optimize operations or create products that were previously too costly to build.

That would mark a major transition for Wall Street.

The first AI boom was largely about financing the machines.

The next phase may be about identifying who uses those machines most effectively.

For big investors searching for tomorrow's winners, that distinction could become one of the defining themes of the market through the rest of 2026.