The Massive AI Opportunity No One is Talking About

If you’ve attended a technology conference, read an earnings call transcript, or scrolled through LinkedIn in the past 18 months, you’ve noticed the same thing I have: AI is inescapable. Every platform—and I mean every platform—has added a copilot. Every software vendor has become an “AI company.” A shoe company has even made the pivot.

The noise is frankly deafening. And if you’re in the technology services business, it may be raising some important questions: Are we being disrupted? Are we too late? Is the opportunity passing us by?

I’ve spent considerable time analyzing technology investment patterns (yes, this is what CFOs do for fun)—not just for AI, but across every major technology wave of the past 50 years. And I’ve reached a conclusion the markets don’t quite seem to understand: the services opportunity in AI is massive… and it’s just getting started.

What makes me so sure? Here’s what the data showed me.

The Pattern I Kept Seeing

Every transformational technology follows a remarkably consistent investment pattern. First, capital floods into hardware—the physical infrastructure that makes the new technology possible. Then, as hardware commoditizes, investment shifts to software—the applications and platforms that create differentiated value. Finally, as the ecosystem matures and complexity grows, services become the dominant investment category—implementation, integration, management, and optimization.

Consider the personal computing revolution. In the 1980s, the winners were hardware companies: IBM, Compaq, Apple. By the early 1990s, value had shifted decisively to software—think of behemoths like Microsoft and Oracle.

 By the late 1990s and into the 2000s, services companies—systems integrators, outsourcers, managed services providers—emerged as a rapidly-growing segment of IT spending, reaching over $624 billion globally by 2005. Critically, services kept on expanding long after hardware and software matured. According to Gartner’s 2026 forecast, IT services spending stands at $1.87 trillion—tripling over two decades and representing the single largest category of IT spending, exceeding both software ($1.44 trillion) and data center hardware ($788 billion).

The pattern repeated with the internet on an even bigger scale.

We often remember the dot-com era through the lens of software and websites. People forget that the hardware phase was enormous. Before a single webpage could load at broadband speeds, someone had to lay the fiber. Companies like Level 3, Global Crossing, WorldCom, and Qwest spent tens of billions of dollars burying fiber optic cable across continents and under oceans. The telecom carriers invested heavily in DSL infrastructure and network buildouts.

This physical infrastructure—the literal pipes of the internet—represented the essential foundation for everything that followed.

The servers, routers, and data centers from companies like Sun Microsystems and Cisco added billions more. Only after this infrastructure was in place did software platforms—web servers, e-commerce applications, content management systems—capture the market’s attention.

And ultimately, the largest sustained opportunity emerged in services: web development, digital marketing, hosting, and managed services. The vast dollar amounts invested over the internet wave split across all three categories, but services had the longest tail.

Cloud computing followed the same trajectory. Amazon Web Services launched in 2006, and the first decade of cloud was dominated by infrastructure buildout, both by hyperscalers constructing data centers and by hardware suppliers feeding that demand. The 2010s saw software take center stage, with the rise of SaaS, PaaS, and cloud-native platforms.

Today, cloud is firmly in the services-dominant phase. Cloud migration, cloud managed services, FinOps, and cloud optimization represent the largest growth categories in enterprise IT spending, with Gartner projecting the public cloud services market will nearly double to $1.42 trillion by 2029.

Where We Are in the AI Cycle—and What Comes Next

So where does AI fit in this pattern? The data tells a clear story: the hardware phase is still running strong; we’re in the early-to-middle stages of the software phase; the services phase is just beginning.

The current AI boom has been extraordinarily hardware-intensive. NVIDIA’s market capitalization has increased by over $2 trillion since early 2023, driven by insatiable demand for GPUs. Hyperscalers are spending hundreds of billions annually on AI infrastructure. Startups are raising massive rounds simply to purchase compute capacity. Anthropic recently secured major commitments for additional compute infrastructure. And Elon Musk’s xAI built Colossus, what they’re calling the world’s largest AI training cluster.

Simultaneously, we’re witnessing explosive growth in AI software. Foundation models from OpenAI, Anthropic, Google, and others are rapidly maturing. AI development platforms, MLOps tools, and enterprise AI applications are proliferating.

All the current noise obscures something huge: the services opportunity in AI will likely be the largest of any technology wave in history. Full stop.

Why? Several factors suggest AI services will over-index relative to historical patterns:

Continuous management. Unlike previous technology waves where you could implement once and maintain in a steady state, AI models require ongoing monitoring, retraining, fine-tuning, and optimization. Model drift is real. Data pipelines change. Business requirements evolve.

This isn’t a one-time implementation; it’s an ongoing operational discipline that enterprises will need partners to manage. This fundamental difference means services won’t just be a phase, but a permanent, expanding category.

Complexity. Implementing AI effectively requires integration with existing systems, data pipelines, security frameworks, and business processes. This is not plug-and-play technology.

Horizontal applicability. Unlike previous waves that primarily affected IT departments, AI touches every function—finance, operations, customer service, HR, legal. The implementation surface area is enormous.

Talent scarcity. The skills required to implement and manage AI effectively are in short supply, with 46% of leaders citing skill gaps as a major barrier to adoption. Enterprises will rely on services partners to fill the gap.

Governance and compliance. Regulatory frameworks are evolving at breakneck speed. Enterprises need help navigating responsible AI practices, bias mitigation, and compliance requirements.

The services opportunity mirrors broader AI expansion as enterprises move from experimentation to scaled production deployments across every business function.

The FinOps Parallel: Why AI Cost Optimization Will Be Essential

Something else that isn’t getting enough attention yet? AI cost management is going to become a critical discipline and a significant services opportunity.

Again, let’s rewind the tape. When enterprises first migrated to public cloud, the promise was efficiency and cost savings. The reality, for many, was runaway spend. Without proper governance, visibility, and optimization practices, cloud bills spiraled out of control.

This gave rise to an entire category of FinOps—financial operations for cloud—encompassing tools, practices, and services to help organizations understand and manage their cloud consumption.

AI is heading down the same path, but potentially faster and with less visibility. Every API call to a large language model consumes tokens. Every inference request has a cost. And unlike cloud infrastructure, where compute costs are relatively well understood, token-based pricing is new, variable, and often opaque.

Enterprises are already experimenting with AI across dozens of use cases—customer service chatbots, document processing, code generation, data analysis—often without full visibility into what they’re spending. If a company as sophisticated as Uber—with world-class engineering and financial operations—can run through its entire annual AI budget just four months into the year, imagine the challenges facing the typical enterprise deploying AI without centralized oversight?

As a CFO with one ear to the ground and the other on calls with clients, I can tell you this is a problem that will demand attention. And it’s not getting easier. While everyone assumes token costs will decline over time, we’re seeing evidence that costs may actually increase as models become more sophisticated and enterprises demand higher quality outputs.

The enterprises deploying AI at scale will need robust governance frameworks, cost allocation models, usage monitoring, and optimization strategies. They’ll need help answering questions like: Which use cases are delivering ROI? Where are we over-consuming? How do we forecast AI spend? How do we optimize prompt engineering and model selection to reduce costs without sacrificing quality?

This represents a substantial services opportunity—AI FinOps—that will parallel what we’ve seen in cloud, but with even greater complexity given the rapid pace of model evolution and the diversity of pricing structures across providers.

Services Leaders, Take Note

For those of us leading technology services businesses, this analysis has several important implications.

First, timing matters—and the timing is favorable. Repeat after me: We are not late to AI; we’re early to the services phase. The enterprises that have been experimenting with AI proofs-of-concept are now asking harder questions: How do we scale this? How do we integrate it with our existing infrastructure? How do we operate it reliably? How do we govern it responsibly? How do we manage the costs?

These are services questions.

Second, our existing capabilities are more relevant than they might appear. Managed services providers have spent decades building expertise in infrastructure management, operational excellence, and enterprise IT complexity. Far from eliminating the need for these capabilities, AI amplifies them.

Models need infrastructure. Platforms need integration. Deployments need ongoing management. Costs need governance. The services companies that can bridge AI innovation with operational reality will capture disproportionate value.

Third, the opportunity requires intentional investment. History shows that services phases reward companies that build relevant practices early. The systems integrators that built cloud practices in 2012-2015 are among those dominating cloud services today. The same will be true for AI.

Services companies that are building AI implementation capabilities, MLOps expertise, AI governance practices, and AI cost optimization frameworks now will be positioned to capture the wave as it accelerates.

Even the AI Companies See it Coming

Lest you think these are just the isolated musings of an optimistic tech services CFO, OpenAI and Anthropic just put their money where my thesis is.

OpenAI launched DeployCo—a consulting and services business—in May 2026, with $4 billion in backing at a $10 billion valuation. Among the investors? Bain, Capgemini, and McKinsey. Anthropic has made similar moves. Goldman Sachs is backing both.

The companies building the foundation models are investing billions in services. Why? Because they see what I see: the model is only valuable if enterprises can actually deploy it, integrate it, and operate it. And that’s a services problem.

When the AI companies themselves are putting billions behind services infrastructure, it tells you something about where this wave is headed.

Cutting Through the Noise

The AI noise in today’s market is real, and it can be distracting. Valuations for AI software companies have reached levels that strain credulity. The hype cycle is operating at full tilt. It’s easy to feel like the opportunity has already been captured by others.

But the pattern of technology investment tells a different story. Hardware phases are exciting and generate headlines. Software phases create iconic companies. But services phases generate the most sustained, broadly distributed value creation—and they last the longest.

For managed services providers, the AI opportunity isn’t behind us. The foundation is just being laid. The enterprises investing in AI infrastructure and experimenting with AI software today will need partners to help them implement, integrate, manage, optimize, and govern tomorrow.

Noise? No. That’s signal. And the moment to respond to it is now.

---

Frequently Asked Questions:

What is the enterprise AI services opportunity?

The enterprise AI services opportunity refers to the rapidly growing demand for implementation, integration, managed services, optimization, and governance support as enterprises move from AI experimentation to scaled production deployments. Following the historical pattern of previous technology waves—personal computing, the internet, and cloud—the services phase represents the largest and longest-lasting investment category. For AI, this phase is just beginning.

What is AI FinOps?

AI FinOps is the emerging discipline of financial operations for AI. It encompasses the tools, practices, and services that help enterprises manage their AI consumption costs—including monitoring token usage and inference costs, allocating AI spend across business units, measuring ROI by use case, and optimizing model selection and prompt engineering to reduce costs without sacrificing output quality. Similar to cloud FinOps, AI FinOps is expected to become a critical enterprise capability as AI spending scales.

Are managed services providers too late to the AI opportunity?

No. The hardware phase of AI investment is still running strong, and the AI software phase is in its early-to-middle stages. The enterprise AI services phase—where managed services providers deliver the most sustained value—is just beginning. Enterprises that have run AI proofs-of-concept are now asking the implementation, integration, governance, and cost management questions that require experienced services partners.

What makes the AI services opportunity larger than previous technology waves?

Three factors suggest AI services will over-index relative to historical patterns: (1) continuous management requirements—AI models require ongoing monitoring, retraining, and optimization rather than one-time implementation; (2) horizontal applicability across every business function, not just IT departments; and (3) severe talent scarcity that forces enterprises to rely on external AI managed services partners for skills they can’t build internally at the pace required.

What should technology services companies do to capitalize on the AI services market?

Technology services companies should invest now in AI implementation capabilities, MLOps expertise, AI governance frameworks, and AI FinOps practices. The historical precedent is clear: the services companies that built relevant practices during the early phase of a technology wave captured disproportionate value. The systems integrators that built cloud practices in 2012–2015 now dominate cloud services. The same pattern will hold for enterprise AI services.

How will AI change the managed services market?

AI will expand the managed services market rather than disrupt it. AI models require ongoing monitoring, retraining, and optimization—creating a permanent, expanding services category rather than a one-time implementation engagement. The integration complexity of enterprise AI deployments, the talent scarcity in AI skills, evolving governance requirements, and the growing need for AI cost management all create sustained demand for managed services providers with deep AI expertise.