Every technological wave brings with it the same familiar prediction – "the incumbents are finished!"

You can hear it in every cycle — the confidence that this time will be different, that this wave will finally collapse the old foundations. Mainframes gave way to minicomputers, minicomputers to PCs, PCs to the internet, the internet to cloud, cloud to mobile, and now mobile to AI. Each shift created a moment where everything felt newly possible, where the ground seemed loose, where young companies assumed the advantage was on their side.

And in those moments, it's easy for founders to look at incumbents and see only decay — old systems, slow roadmaps, thick layers of process, bloated product suites. A sense that nothing inside is truly moving.

But when you spend real time in the legacy economy — inside the operations, systems, and workflows that quietly hold entire industries together — the picture looks different.

I spent years in these environments, tracing the logic of industries through their systems, their people, and the invisible decisions that shaped them. Freight networks stitched together by tribal knowledge. Medical billing systems powered by thousands of micro-decisions. HVAC service operations juggling paper logs and brittle scheduling tools. Compliance-heavy businesses where ritual mattered as much as software. Industries dominated by private equity, shaped by regulation, and held together by tools that were rough on the surface but precise in the ways that actually mattered.

To outsiders, these ecosystems looked vulnerable. To insiders, they looked inevitable.

Incumbents didn't dominate because their products were elegant. They dominated because they had become part of the industry's identity. Every workflow, every report, every integration, every operator's muscle memory had wrapped itself around the incumbent's logic. What looked like "slowness" was often the residue of tens of thousands of small choices that had hardened over decades.

This is the part of incumbency most founders underestimate: the historical weight embedded in the system.

— Yesterday's decisions becoming today's constraints.

— Switching costs that stretch far beyond software.

— Regulatory structures that make change expensive.

— Multi-system dependencies where altering one layer destabilizes five others.

— Risk asymmetry where no one is punished for choosing the old vendor, but many have been punished for choosing the new.

— Data and workflows so interwoven that replacing them feels like surgery.

These forces aren't abstract. They are the moats that keep industries stable.

And when you're building something new, it's tempting to see all of this as opportunity — as incompetence waiting to be disrupted. But the truth is more nuanced: what looks inefficient is often the only configuration the industry can tolerate.

I remember sitting with an operator at a logistics company early on, watching her navigate a 27-step process across three disconnected systems. The keystrokes, the tabs, the refreshes, the cross-checks; it felt absurd. But when I asked why they hadn't upgraded to something simpler, she didn't hesitate:

"We tried that once. It broke everything."

That sentence captured years of institutional memory. It captured why incumbents endure. And it captured what most founders miss when they dismiss them.

This is why incumbents appear sleepy: not because they don't care, but because they carry systems fragile enough that moving recklessly would break the businesses they serve. Their pace is shaped by the complexity of the industry, the expectations of their customers, and the constraints written into their architecture.

But the same forces that make incumbents durable also make them vulnerable — particularly when the substrate of the industry begins to shift.

And that is what makes the AI era so interesting.

AI challenges incumbents in a way previous waves couldn't. It doesn't compete at the interface layer or the feature layer. It competes at the workflow layer — collapsing processes incumbents were built around, automating judgment that once required human review, and dissolving the activities that shaped incumbents' value propositions.

Classification, reconciliation, exception handling, routing, coordination — tasks that formed the backbone of entire industries — suddenly become cheap, near-instant, or unnecessary.

This is why so many founders feel like the field is open again, that AI-native systems can finally step in and rewrite the rules. And some will. The early signals are real.

But the part that gets overlooked is this: incumbents may look sleepy, but they are not powerless. They remain the gravitational center of their industries. They still own the relationships, the data, the integrations, the compliance position, the distribution channels, and the trust customers rely on for continuity.

A young company may be able to ship faster. But an incumbent, once awake, can shift an entire sector by moving one layer closer to the new paradigm.

What most founders mistake for weakness is often stored potential — the kind that compounds quietly and reveals itself suddenly.

In an AI-native era, with the ground feeling loose and possibility everywhere, it's tempting to assume incumbents will finally give way. But if history teaches anything, it's this: the companies that win are the companies that understand the forces they're up against.

Some founders study incumbents. Most don't. And the difference is everything.

When you ignore an incumbent, you're ignoring the architecture of the industry. When you don't understand their workflows, you don't understand the customer. When you underestimate their distribution, you underestimate your own timeline. When you miss their constraints, you overestimate your speed. When you don't study their past, you misread your future.

The industries I spent time in taught me that every market tells its story through the incumbents who shaped it. To build successfully inside that industry, you need to understand that story with precision.

In AI, the stakes are even higher. Because the incumbents' sleep may be deeper — but their awakening may be faster.

A sleeping giant is still a giant. And if you want to build in its shadow, you need to know the shape of its posture, the history of its reflexes, the structure of its habits, and the boundaries of its strength.

The future will reward founders who understand incumbents not as obstacles but as artifacts of the market — artifacts that explain how the industry works, where the pressure points live, and where new systems might find real leverage.

The AI era feels wide open. But openness is not the same thing as emptiness. The ground you're standing on was shaped long before you arrived.

If you want to compete, you don't start by ignoring incumbents. You start by studying them.

Because the companies that appear sleepy are often the ones that have learned how to sleep with one eye open.