Every few decades, a technology arrives that separates companies into two groups: those that built with it early, and those that got replaced by someone who did.
We are in one of those moments right now. And most companies are making the same mistake.
The Pattern
Between 1975 and 1995, the hard disk drive industry went through one of the most studied cycles of disruption in business history. Clayton Christensen documented the whole thing in The Innovator’s Dilemma, and the story he tells is not about bad companies making bad decisions. It is about good companies, well-managed companies, making decisions that were perfectly rational at every step and still losing everything.
In the late 1970s, the market was dominated by makers of 14-inch disk drives. These were serious firms, deep engineering talent, strong customer relationships, selling to mainframe manufacturers who trusted them. Then a new class of drive appeared: the 8-inch drive. It was smaller. It held less data. It was slower. By every metric that mattered to mainframe buyers, it was worse.
The 14-inch drive companies did what any responsible company would do. They asked their customers whether they wanted this new, smaller drive. Their customers said no. They wanted more capacity, more speed, better cost per megabyte. The 8-inch drive couldn’t compete on any of those dimensions. So the incumbents kept investing in what their customers wanted, and ignored the 8-inch form factor.
But the 8-inch drive found a different market: minicomputers. A market the mainframe drive makers weren’t serving and didn’t care about. In that market, the 8-inch drive was good enough. And over time it improved. More capacity. Better speed. Lower cost. Eventually it became good enough for mainframes too. By the time the 14-inch makers realized what had happened, it was too late. The 8-inch companies had the scale, the manufacturing expertise, and the customer relationships. The 14-inch makers were finished.
Here is the part that should keep every executive awake at night: this same pattern repeated. The 8-inch companies became dominant. Then the 5.25-inch drive appeared. The 8-inch companies asked their minicomputer customers if they wanted it. The customers said no. They wanted better performance on the metrics they already cared about. So the 8-inch companies ignored the 5.25-inch drive. It found a home in desktop PCs instead. And the cycle repeated. 5.25-inch gave way to 3.5-inch. 3.5-inch to 2.5-inch. Every time, the established leaders talked to their customers, heard “no,” and made the rational decision to stay the course.
Every single generation of leaders watched the previous generation get destroyed by this exact dynamic. And they fell into the same trap anyway. Not because they were stupid. Because the logic of “build what your best customers are asking for” is so compelling that it overrides everything else.
The AI Version of This Story
Right now, most established companies are doing one of two things with AI: automating existing operations to reduce costs, or bolting AI features onto existing products to stay competitive. Both feel productive. Both have clear ROI. Both are the equivalent of building faster 14-inch disk drives.
Automation is a margin play. It makes the existing business cheaper to run. That is valuable, but it has a ceiling. No company has ever automated its way to growth.
AI features on existing products are a retention play. They keep customers from leaving. Also valuable, also limited. You are running to stay in place. And every feature you bolt onto a legacy architecture inherits that architecture’s cost, complexity, and constraints. You are adding capability while also adding weight.
Neither of these is a growth strategy. Neither changes your competitive position. Neither builds the thing that will define your market in five years.
The Real Investment
The companies that will lead the next era are not the ones spending the most on AI. They are the ones building new products on new architectures, products that could not have existed before this technology, products that are not burdened by the decisions and infrastructure of the previous generation.
This is the hard one. It does not have a clear ROI spreadsheet. There is no client asking for it. There is no demand signal. The business case is: “if we don’t build this, someone else will, and they won’t be carrying our technical debt when they do.”
That is not a case most planning processes are designed to approve.
The Comfortable Trap
The trap is that everything you are already doing feels responsible. Automation has measurable savings. Client-requested features have a name attached to the request. AI on your existing platform ships faster because the infrastructure is already there.
Building something genuinely new, on a modern foundation, without a guaranteed customer? That feels like a risk. It feels speculative. It feels like the kind of thing you do after you finish the important work.
This is exactly how every displaced incumbent felt. The important work was so important that the transformative work never got funded. Until a competitor funded it instead.
The Window
Here is the thing about being an incumbent: you have advantages a startup does not have. You have domain expertise, customer relationships, regulatory knowledge, operational credibility, and distribution. These are real advantages, and they are depreciating.
Domain expertise can be learned. Customer relationships can be built. Regulatory knowledge can be hired. Every month you spend optimizing only your existing business is a month in which a new entrant is closing the gap on everything you think protects you.
There is a window during which you can combine your existing advantages with the new technology. That window is open now. It will not stay open. And “we’ll get to it after we finish our current priorities” is exactly the sentence that has preceded every disruption in business history.
What To Do About It
This is not a call to abandon your existing business. That business pays the bills, keeps your promises, and funds everything else. Keep running it. Keep improving it.
But carve out a real effort, with real resources, real people, and a real mandate, to build something new. Not new features on old architecture. New products on new foundations. Give that effort different success metrics. Do not force it to compete for priority against work that has clearer short-term returns. Protect it from the gravitational pull of the existing roadmap.
The companies that survived previous disruptions were the ones that ran both: the sustaining business and the disruptive one. In parallel. With the understanding that the second one might look unimpressive for a while, and that this was fine, because the goal was not immediate revenue. The goal was still being relevant in five years.
The Choice
You are not behind yet. That is the good news.
The bad news is that “not behind yet” is a time-limited condition. The future of your industry is getting built right now. With you, or without you.