‍

Everyone is asking the same question:

‍

Where is the next big AI opportunity beyond semiconductors, AI datacenters, and foundational models?

‍

At the same time, enterprises around the world have been aggressively investing in AI. OpenAI, Anthropic, Microsoft, Google, and others have spent the last two years selling AI capabilities to large organizations with the promise of productivity gains, automation, and new growth opportunities.

‍

Yet the return on those investments remains uncertain for many companies.

‍

Multiple studies point to this gap. A 2024 Deloitte survey found that while most organizations are increasing AI spending, only a small percentage report achieving significant financial benefits from their generative AI initiatives. Similarly, a 2024 IBM study found that many executives expect AI to drive value, but relatively few organizations have successfully scaled AI deployments that produce measurable business outcomes. McKinsey’s research has also shown that while AI adoption continues to rise, only a minority of companies report material impacts on earnings from AI initiatives.

‍

My view is that we’re still looking at the wrong part of the stack.

‍

We Are Still Building the Infrastructure

‍

Today, most of the activity is focused on building the foundation: semiconductors, AI datacenters, networking, power, cooling, and large language models.

‍

This is essential work.

‍

Every major technology revolution starts by building the infrastructure first.

‍

Railroads came before modern commerce. The internet came before Google. Smartphones came before Uber.

‍

Infrastructure creates possibilities.

‍

And that is exactly where we are today.

‍

While enterprises are experimenting with AI and vendors are racing to sell AI solutions, the industry is still heavily focused on building the underlying infrastructure that makes large-scale AI possible.

‍

Historically, however, the largest number of new companies, new markets, and new business models emerge after the infrastructure is in place.

‍

That doesn’t mean infrastructure won’t remain valuable. It will. In fact, AI may be more infrastructure-intensive than any technology wave we’ve seen before.

‍

But I don’t believe the biggest opportunity over the next decade lies solely in building better chips, larger datacenters, or bigger models.

‍

I believe the bigger opportunity lies in what gets built on top of them.

‍

And that distinction helps explain why so many organizations are struggling to generate meaningful returns from their AI investments.

‍

Why Many Enterprises Aren’t Seeing ROI

‍

Many enterprises are spending heavily on AI and wondering why the returns haven’t materialized.

‍

I don’t think the problem is AI.

‍

I think the problem is how organizations are approaching it.

‍

Most companies are asking:

‍

“How do we add AI to what we already do?”

‍

The better question is:

‍

“If AI existed from day one, how would we design this process, product, or business differently?”

‍

That is a fundamentally different conversation.

‍

Many organizations are using AI to improve existing workflows by 10% or 20%.

‍

Few are reimagining the workflow itself.

‍

As a result, they get incremental efficiency gains when they are hoping for transformational outcomes.

‍

This helps explain why so many enterprise AI deployments have produced mixed results. Companies often purchase AI tools, launch pilots, and deploy copilots, but the underlying business process remains largely unchanged.

‍

The challenge is no longer access to AI technology.

‍

The challenge is identifying the right problems to solve.

‍

The companies that win won’t start with AI.

‍

They’ll start with the customer problem, the business problem, or the market opportunity and then determine how AI can create a dramatically better outcome.

‍

The Next Wave Is AI-Native Applications

‍

Most of what we see today is AI-enabled software.

‍

Existing products are adding copilots, assistants, chat interfaces, and automation capabilities.

‍

That’s valuable.

‍

But I believe the bigger opportunity lies in AI-native applications.

‍

The distinction matters.

‍

AI-enabled products improve existing workflows.

‍

AI-native products create entirely new workflows.

‍

They assume from day one that intelligence is abundant, software can reason, agents can act autonomously, and knowledge work can be performed at dramatically lower cost.

‍

That changes everything.

‍

It changes user experience.

‍

It changes organizational design.

‍

It changes operating costs.

‍

It changes how decisions are made.

‍

And it changes what products can exist in the first place.

‍

We’re still in the very early stages of this transition.

‍

AI Will Rewrite Business Models

‍

The opportunity is not just about applications.

‍

It is also about business models.

‍

One of the most important consequences of AI is that it dramatically lowers the cost and complexity of creating new products and services. When building becomes easier, markets can move from standardized, mass-produced offerings toward highly tailored solutions designed for specific customers, industries, or even individuals.

‍

Software is the clearest example.

‍

Historically, software companies built one product and sold it to thousands or millions of customers. Every customer adapted their workflow to fit the software.

‍

AI changes that equation.

‍

Instead of forcing customers into the same product, companies can increasingly create software that adapts itself to each customer’s processes, preferences, data, and objectives. The economics of customization improve dramatically when intelligence becomes abundant.

‍

But this shift extends far beyond software.

‍

In consumer markets, we may move from mass-market products to personalized experiences. Retailers can create individualized shopping journeys, customized product bundles, personalized education programs, nutrition plans, travel itineraries, and entertainment experiences tailored to each customer.

‍

In industrial markets, manufacturers can move beyond standardized equipment toward highly customized systems optimized for a specific factory, production line, or operating environment. Engineering, design, maintenance, and operational workflows can be configured dynamically rather than delivered as one-size-fits-all solutions.

‍

In financial services, institutions can move beyond generic products such as standard loans, portfolios, and insurance policies toward solutions tailored to an individual business’s financial goals, risk profile, cash flow patterns, and life circumstances. Financial products become increasingly personalized rather than broadly segmented.

‍

In professional services, firms have historically relied on expensive human expertise to deliver customized outcomes. AI makes it possible to scale expertise itself. Legal, consulting, accounting, marketing, and advisory services can become far more tailored while remaining economically viable for a much larger customer base.

‍

This shift has profound implications for business models.

‍

For decades, scale came from standardization. Companies created value by selling the same product to as many customers as possible.

‍

AI enables a different model: scale through customization.

‍

As products become more adaptive and outcome-oriented, customers will increasingly pay for results rather than access.

‍

A recruiter may be paid per successful hire.

‍

A legal platform may be paid per completed matter.

‍

A customer service platform may be paid per issue resolved.

‍

A sales platform may be paid per qualified opportunity generated.

‍

When products deliver customized outcomes rather than standardized capabilities, pricing models, customer relationships, organizational structures, and competitive advantages all begin to change.

‍

That is a much larger shift than many people realize.

‍

Entirely New Companies Will Emerge

‍

Every major technology wave creates businesses that were previously impossible.

‍

Uber wasn’t simply a better taxi company.

‍

It became possible because GPS, smartphones, and ubiquitous internet access reached maturity at the same time.

‍

Those technologies enabled an entirely new business model.

‍

AI will have a similar effect.

‍

AI dramatically reduces the cost of cognition.

‍

For the first time, expertise can be replicated, scaled, personalized, and delivered at near-zero marginal cost.

‍

That will create opportunities across healthcare, education, financial services, logistics, manufacturing, and professional services.

‍

Some of the most important companies of the next twenty years may not exist today.

‍

Just as few people predicted Google, Facebook, Airbnb, or Uber at the beginning of earlier technology cycles, many of the defining AI companies are likely still being conceived.

‍

The biggest winners may not be the companies adding AI features to existing products.

‍

They may be companies built from the ground up for an AI-native world.

‍

The Leadership Gap

‍

One observation stands out to me.

‍

Technology is becoming increasingly accessible.

‍

Capital is available.

‍

Models are becoming commoditized.

‍

The scarce resource is no longer technology.

‍

The scarce resource is leadership imagination.

‍

The leaders who create the most value won’t necessarily be the ones with the largest AI budgets.

‍

They will be the ones who can envision entirely new products, workflows, business models, and markets.

‍

Many organizations are still treating AI as a technology initiative.

‍

The best organizations are treating it as a business reinvention opportunity.

‍

That distinction will matter.

‍

Build or Join

‍

My advice is simple.

‍

Don’t wait for established companies to figure out what AI means for your career.

‍

Figure out what AI means for your own capabilities and opportunities.

‍

That might mean building a company.

‍

It might mean joining an AI-native startup.

‍

It might mean partnering with founders who are creating new categories.

‍

But the biggest opportunities rarely appear once everyone agrees they exist.

‍

By then, much of the value has already been created.

‍

Final Thought

‍

Today, most of the attention is focused on chips, datacenters, and foundational models.

‍

Those investments are necessary.

‍

But they are laying the groundwork for something larger.

‍

The next decade will not be defined solely by infrastructure.

‍

It will be defined by the applications, business models, and companies that infrastructure makes possible.

‍

We’re still building the roads.

‍

At the same time, enterprises are trying to figure out how to generate meaningful returns from AI investments that have yet to consistently deliver on their promise.

‍

That gap between infrastructure and value creation is where the next wave of opportunity will emerge.

‍

The real question is:

‍

Who will build the next generation of businesses that travel on these roads?