- media@moneyinfocus.news
For much of the past three decades, every major technological revolution has followed a familiar pattern. The internet rewarded connectivity. Mobile computing rewarded scale. Cloud computing rewarded recurring software revenue.
Artificial intelligence is following a different script.
Instead of beginning with consumer applications or enterprise software, the current cycle has started with something far less visible—but far more capital intensive. Before AI can transform industries, companies must first build the infrastructure capable of supporting it. That has created what many market observers describe as the largest capital deployment cycle in the history of the technology sector.
According to the Stanford AI Index Report 2026, private investment in artificial intelligence remains at historically elevated levels despite a moderation from the peak funding environment of previous years. Meanwhile, adoption among large enterprises continues to accelerate, with generative AI becoming a strategic priority across finance, healthcare, manufacturing, defense and professional services.
The investment landscape tells an even larger story.
Research from McKinsey & Company estimates that generative AI could ultimately generate between US$2.6 trillion and US$4.4 trillion in annual economic value, fundamentally reshaping productivity across virtually every sector of the global economy. PwC projects that artificial intelligence could contribute nearly US$15.7 trillion to global GDP by 2030, making it one of the most significant economic transformations since electrification. Wall Street has responded accordingly.
Technology companies are committing hundreds of billions of dollars toward data centers, semiconductor manufacturing, networking equipment, cloud infrastructure and energy generation—not because those assets directly produce intelligence, but because they represent the foundation upon which tomorrow’s AI economy will be built.
Now, the question is no longer whether AI will transform business. The question is who will capture the economic value.
That distinction matters because history suggests technological revolutions rarely reward every participant equally. During the dot-com era, thousands of internet companies disappeared while a handful became trillion-dollar enterprises.
The cloud revolution created dominant software franchises while rendering countless legacy vendors obsolete. Artificial intelligence is expected to produce a similarly uneven outcome.
Paul Rodriguez, Senior Technology Analyst at R.F. Lafferty & Co., believes the current cycle differs from previous technology booms in one fundamental respect.
“AI, its benefits, and risks have taken longer to assess than past booms, but the implementers of the technology have not slowed down as they see the benefits.”
His observation captures one of today’s defining market paradoxes: while governments continue debating regulation, ethics and governance, capital spending has accelerated rather than slowed. Investors have accepted that uncertainty exists—but they appear increasingly convinced that waiting carries greater strategic risk than investing.
Artificial intelligence has become one of the largest capital expenditure stories ever witnessed in global financial markets.
Unlike previous software cycles, where innovation primarily required engineers and venture funding, today’s AI leaders require extraordinary physical infrastructure.
Modern frontier models demand enormous computational resources, specialized chips, advanced cooling systems, high-speed networking, reliable electricity and vast quantities of water.
Every new generation of AI models increases those requirements. As a result, Wall Street’s attention has shifted away from software alone and toward the industries enabling AI at scale.
Few sectors illustrate this better than semiconductors.
NVIDIA has emerged as the defining infrastructure company of the AI era, with its graphics processing units becoming the preferred hardware for training and deploying large language models. Its rapid revenue growth has reflected unprecedented demand from hyperscale cloud providers racing to expand AI capacity.
Yet semiconductors represent only one layer of a much larger investment story.
Microsoft has committed tens of billions of dollars toward expanding Azure’s AI infrastructure while deepening its partnership with OpenAI.
Amazon continues investing aggressively in AWS data centers and proprietary AI chips.
Alphabet is simultaneously expanding Google Cloud while developing custom Tensor Processing Units for AI workloads. Meta has significantly increased capital expenditures to support open-source foundation models, recommendation systems and next-generation AI infrastructure.
Oracle has also repositioned itself as an infrastructure provider, investing heavily in cloud computing capacity to support enterprise AI deployment.
Meanwhile, OpenAI, xAI and Anthropic are collectively driving one of the largest infrastructure procurement cycles ever seen among privately held technology companies. Their demand for computing power has reshaped supply chains spanning semiconductors, networking equipment, memory manufacturers and data center operators.
The scale of spending has surprised even seasoned investors. Rather than competing solely through algorithms, AI companies increasingly compete through access to compute.
That reality has fundamentally changed how institutional investors evaluate the sector.
According to Rodriguez, capital is not concentrating in one isolated corner of artificial intelligence. Instead, investment is flowing throughout the ecosystem.
“All the above, but I would also include energy, water, and computing hardware including memory.”
That perspective reflects an increasingly important shift in institutional thinking: while much of the public conversation focuses on chatbots and generative AI applications, professional investors are evaluating the underlying industrial economy required to sustain those technologies.
In many respects, artificial intelligence resembles previous industrial revolutions more than previous software revolutions. Just as railroads required steel, electricity required transmission networks, and cloud computing required hyperscale data centers, AI requires an entirely new physical backbone capable of supporting unprecedented computational demand.
One characteristic distinguishes the current investment cycle from nearly every previous technology boom: capital allocation is occurring before widespread monetization.
Historically, markets rewarded companies after new technologies demonstrated durable revenue generation.
Artificial intelligence has largely reversed that sequence. Investors are funding infrastructure first while expecting applications and profits to emerge later.
That helps explain why hyperscalers continue raising capital expenditure forecasts despite ongoing questions regarding near-term returns.
According to Goldman Sachs Research, AI-related infrastructure spending could exceed US$1 trillion over the coming years, driven primarily by hyperscale cloud providers and enterprise demand for computing capacity.
BlackRock has similarly described artificial intelligence as one of the defining long-term structural investment themes reshaping global capital allocation, emphasizing that infrastructure—not merely software—will likely capture a substantial share of future investment flows.
Rodriguez argues that investors should avoid interpreting this spending solely as speculative enthusiasm.
Instead, Wall Street is increasingly viewing AI through the lens of operational efficiency. “Wall Street is pricing AI as an efficiency solution allowing users to deploy technology faster with a less expensive cost attached.”
However, he also cautions that investors may be underestimating an important variable. “The assumptions that are overly optimistic are the costs associated with the risk of deploying AI,” he notes.
That distinction may ultimately determine which companies justify today’s elevated valuations. The race is no longer simply about building the most capable AI model.
It is increasingly about building the most economically sustainable one.
As Wall Street continues directing unprecedented amounts of capital toward artificial intelligence, the industry’s next phase may depend less on technological breakthroughs than on whether companies can convert extraordinary infrastructure investment into durable competitive advantage and long-term cash flow.
The first wave of artificial intelligence investing has been defined by one dominant theme: infrastructure. Chips, cloud platforms, networking equipment, hyperscale data centers and power generation have attracted hundreds of billions of dollars in capital as companies race to build the foundation of the AI economy.
But every technology cycle eventually reaches the same turning point.
The market stops asking who is building the technology and starts asking who can actually make money from it. That transition is already beginning.
Although AI remains one of the strongest structural investment themes globally, institutional investors are becoming increasingly selective. Rather than rewarding every company that mentions artificial intelligence, professional investors are placing greater emphasis on execution, capital efficiency, monetization and sustainable returns.
For Paul Rodriguez, Senior Technology Analyst at R.F. Lafferty & Co., this evolution is both natural and healthy.
“The market has moved beyond asking whether AI matters,” he says. “The next phase is determining which companies can transform massive investment into durable business models.”
Perhaps no question has dominated Wall Street more than whether artificial intelligence has entered speculative territory.
The comparison with the dot-com era is almost inevitable. Technology valuations have expanded rapidly, private AI startups continue raising funding at extraordinary valuations, and infrastructure spending is accelerating at a pace rarely seen in modern capital markets.
Yet Rodriguez does not believe today’s environment resembles the excesses of the late 1990s.
Money In Focus: Do you believe parts of the AI market are already entering bubble territory?
Paul Rodriguez: “I believe that AI markets are not in bubble territory given where we are in the deployment cycle.”
His reasoning reflects how institutional investors distinguish enthusiasm from speculation. Unlike the internet boom, today’s investment cycle is being driven primarily by some of the world’s most profitable companies. Microsoft, Alphabet, Amazon and Meta are funding AI expansion through substantial operating cash flows rather than speculative financing.
The infrastructure being built today is also generating immediate enterprise demand. Businesses across financial services, healthcare, cybersecurity, manufacturing and professional services are already deploying generative AI tools to improve productivity and automate workflows.
That does not eliminate risk. It simply changes where investors should look for it.
Public attention often focuses on product launches and chatbot capabilities. Institutional investors focus elsewhere. They examine whether companies are producing measurable returns on the enormous amounts of capital being deployed.
Money In Focus: Which indicators do you personally monitor to distinguish sustainable innovation from speculative excess?
Paul Rodriguez: “Indicators I look at are CapEx growth rates, top-line growth rates, pricing of various compute inputs like memory, storage, and chip prices.”
Those metrics provide a more complete picture than share price appreciation alone. Capital expenditure reveals how aggressively companies are investing.
Revenue growth indicates whether demand is keeping pace.
The pricing of memory, storage and compute infrastructure reflects whether supply chains remain constrained or begin normalizing.
For professional investors, these variables often matter more than headline announcements surrounding the latest AI model.
Much of the investment narrative surrounding artificial intelligence has centered on semiconductors.
Companies manufacturing GPUs have become synonymous with the AI revolution, and for good reason. Without accelerated computing, modern foundation models would not exist.
Rodriguez agrees semiconductors remain indispensable. He also believes the market may be overlooking the industry’s next major opportunity.
Money In Focus: Which segment of the AI market is receiving excessive attention, and which remains underestimated?
Paul Rodriguez: “Excessive attention is probably given to semiconductors (but this area is critical). I believe that application software has been underestimated as software will be a key component in AI build-out.”
That perspective challenges one of the most common assumptions surrounding generative AI. Many investors fear AI will replace existing enterprise software. Rodriguez sees a different outcome.
Software companies capable of embedding AI into existing workflows may become some of the largest beneficiaries of the entire technology cycle.
Enterprise applications remain the interface through which businesses manage finance, sales, operations, customer relationships and supply chains.
Artificial intelligence is unlikely to eliminate those systems. Instead, it may dramatically increase their value.
While discussions about AI frequently emphasize productivity gains, Rodriguez believes investors continue to underestimate another side of the equation.
Those factors could significantly influence long-term returns.
Money In Focus: What risks surrounding AI adoption are institutional investors currently underestimating?
Paul Rodriguez: “The risks that surround AI adoption and commercialization are costs to deploy and maintain along with expense of securitizing AI.”
As AI systems become embedded in critical infrastructure and enterprise operations, cybersecurity requirements become increasingly complex.
Large language models require continuous monitoring, governance frameworks, secure data environments and regulatory compliance.
Those costs may prove just as important as compute itself.
The first bottleneck of the AI revolution was semiconductor availability. The next one may look very different.
According to Goldman Sachs, electricity demand from data centers is expected to accelerate sharply over the coming decade as AI adoption expands.
Grid modernization, transmission capacity and reliable baseload power are increasingly viewed as strategic assets.
Paul argues the challenge extends well beyond energy.
Money In Focus: Could infrastructure constraints become the defining bottleneck of the next decade?
Paul Rodriguez: “Absolutely, and I would add other items in this list, including water and navigating state governments who in some cases are restricting data center deployments.”
His answer highlights a structural issue that receives relatively little public attention: modern AI data centers consume enormous quantities of electricity for computation and large volumes of water for cooling. They also require local permitting, environmental approvals and access to transmission infrastructure.
The next phase of AI competition may therefore be influenced as much by public policy and physical infrastructure as by advances in machine learning itself.
Technological revolutions create winners. They also expose businesses unable to adapt.
Paul Rodriguez believes the greatest vulnerability does not lie with hardware manufacturers but with software vendors that fail to integrate AI into their products.
Money In Focus: Which business models are most vulnerable during the AI transition?
Paul Rodriguez: “Application software vendors who are not able to figure how to leverage AI are most vulnerable in the AI transition.”
Artificial intelligence has entered a more sophisticated stage of its investment cycle.
Infrastructure remains essential, but investors are increasingly shifting their attention toward execution, economics and long-term sustainability.
The next phase of the AI economy is unlikely to be defined by who builds the most advanced models, but by who consistently translates technological leadership into measurable financial performance. That shift, from expectations to fundamentals, will likely determine the sector’s long-term winners.
—–
Paul Rodriguez is Senior Technology Analyst at R.F. Lafferty & Co., Inc., where he provides independent technology research and investment analysis for institutional investors. With more than 30 years of experience across Wall Street and the software industry, he has held senior research and leadership positions at firms including Needham & Company, CE Unterberg Towbin, Detwiler Fenton, Arthur Wood, and R.F. Lafferty.
