Jim Cramer is urging investors to look more carefully at where the next phase of stock-market gains may come from as Wall Street reassesses some of the biggest beneficiaries of the artificial-intelligence boom.

The Mad Money host's latest market commentary comes at a time when investors have begun rotating between different parts of the technology and growth complex. While AI infrastructure remains one of the market's most powerful themes, some investors are increasingly looking for companies that can benefit from AI spending without carrying the same valuation and expectations associated with the industry's largest names.

Recent commentary surrounding Cramer's views has highlighted optical and photonics companies serving AI data centers, including Lumentum and Coherent. Cramer has argued that these businesses provide another way to participate in the AI infrastructure expansion.

The first phase of the AI investment boom was dominated by companies such as Nvidia, whose processors became essential to the construction of large AI data centers.

As spending has expanded, however, the infrastructure supporting those chips has become increasingly important.

AI servers require high-speed connections capable of moving enormous quantities of data. That creates demand for optical components, transceivers, lasers and other networking technologies.

This is where companies such as Lumentum and Coherent enter the picture.

Their products may not receive the same attention as AI processors, but they are part of the infrastructure required to connect increasingly powerful computing systems.

Cramer's argument effectively broadens the AI investment thesis from semiconductors to the entire data-center supply chain.

Traditional electrical connections face physical limitations as data-center bandwidth requirements rise.

Optical technology can move data rapidly over longer distances and with high bandwidth, making it increasingly important for large-scale AI clusters.

As AI models become larger and companies deploy more accelerators, the amount of information moving between processors also increases.

That means data-center operators cannot simply add more computing power without considering how those processors communicate with each other.

The networking layer becomes a potential bottleneck.

This is one reason Cramer has highlighted photonics companies as a potentially attractive way to participate in the continuing AI infrastructure buildout.

Among the companies discussed in Cramer's recent AI-infrastructure thesis, Lumentum has been one of the stronger performers.

Its fiscal second-quarter revenue reached approximately $665.5 million, up 66% from a year earlier, while non-GAAP earnings per share came in at $1.67. The company's non-GAAP operating margin also expanded significantly.

Those figures illustrate why investors have become increasingly interested in the optical infrastructure supply chain.

The AI boom is no longer simply producing enormous orders for GPUs.

It is creating demand across networking, cooling, power equipment, memory, optical communications and data-center construction.

Another reason investors are paying attention is Nvidia's financial involvement in the optical ecosystem.

Nvidia has invested billions of dollars in companies involved in AI data-center connectivity, reinforcing the view that networking technology will play an increasingly important role as AI clusters scale.

Cramer has pointed to those investments as evidence that the underlying infrastructure opportunity extends beyond the most obvious AI semiconductor names.

That does not guarantee that every company connected to AI infrastructure will succeed.

Valuations still matter, competition remains intense and AI spending could eventually slow.

But it demonstrates how much broader the AI capital cycle has become.

The biggest issue for investors is that the AI trade has become crowded.

Many of the most obvious beneficiaries have experienced enormous gains, creating concerns about valuation and expectations.

That does not necessarily mean the AI boom is ending.

Instead, it could mean that the market is beginning to distinguish between companies simply associated with AI and companies that are actually seeing their revenue and earnings rise because of AI demand.

That distinction could become increasingly important.

Companies with strong backlogs, improving margins and exposure to expanding data-center spending may continue to attract investors even if enthusiasm for some of the most expensive AI stocks fades.

The broader market may therefore be entering a period in which AI remains a dominant theme but leadership becomes more diversified.

Instead of a small group of semiconductor companies capturing most of the gains, investors could increasingly spread their exposure across the infrastructure ecosystem.

That could include optical networking, power management, cooling systems, electrical equipment and data-center construction.

The shift would not necessarily be negative for Nvidia or other major AI-chip companies.

In fact, continued investment throughout the supply chain could reinforce the underlying demand for their products.

The bigger question is whether spending can continue growing quickly enough to justify today's valuations.

For Cramer, the answer appears to involve looking beyond the headline AI names.

The next opportunity may lie in companies that quietly provide the physical infrastructure needed to keep increasingly powerful AI systems running.

That makes the current market less about finding the next Nvidia and more about identifying the businesses that benefit as the AI ecosystem becomes larger, faster and more interconnected.