AI Is Not a Revolution: What Nobody Has Ever Told You

The innovator's dilemma

To understand what is truly happening today, one must return to one of the most important texts ever written on innovation: The Innovator's Dilemma.

Clayton Christensen articulated a concept as simple as it is devastating: leading companies rarely fail because they make the wrong choices. They fail because they keep making the right choices… in a context that has since changed. Disruption never arrives in the form of perfect technology. It enters from below — it appears inferior, costs less, and is more accessible. Incumbents ignore it, because it does not yet meet their standards. But then it improves, evolves rapidly, and, almost without the market noticing, it climbs the value chain until it renders the formerly dominant offering irrelevant.

It is the same pattern that, throughout market history, has already destroyed companies that appeared unassailable. Not because they were surpassed by something superior, but because they were replaced by something cheaper, more accessible, and initially underestimated.

Historical precedents

And this is precisely the pattern now repeating itself with artificial intelligence. Except that, this time, there is one enormous difference: speed.

In all of these cases, the point was not the initial quality of the new technology, but the fact that it was changing the economic rules of the sector. And when the rules change, even market leaders become vulnerable.

The economic dimension of the phenomenon

The numbers help to convey the true scale of the phenomenon. According to McKinsey & Company, generative AI has the potential to impact up to 60–70% of work activities and generate between $2.6 and $4.4 trillion in economic value each year. This is not a technological niche, but a structural shift affecting the way work is performed and, above all, the way value is created and distributed.

But the most important figure is not even this. It is the fact that this technology is spreading faster than the Internet did. Within just a few years, a significant share of the population and of businesses has already begun to incorporate it into daily processes. This means one thing only: the time companies have to adapt has been drastically reduced. And when time shrinks, the probability that someone will be left behind increases.

When scarcity disappears

Imagine a company that builds its business around a replicable product or service: content, customer support, basic software development, standardised analysis, administrative processes. That product carries a price because it requires time, expertise and personnel. Now imagine that artificial intelligence can do the same thing, faster and at a marginal cost approaching zero. At that moment, something irreversible happens: that product ceases to be scarce.

And when something ceases to be scarce, the market is no longer willing to pay for it as it once did.

There is no need for a company to disappear overnight. Companies do not die suddenly. They hollow out. First they lose the ability to charge for their product, then margins begin to erode, growth fades… and by the time the market takes notice, it is too late. The repricing has already begun.

This dynamic is not theoretical. It has already started. In recent months we have witnessed violent sell-offs across several sectors — software in particular — where hundreds of billions in market capitalisation were wiped out within days on the back of a very specific narrative: the risk of AI disruption. And it did not stop there. Services, healthcare, real estate, analytics: the market is beginning to strike, often indiscriminately, at anything that appears even remotely vulnerable.

The market anticipates the fundamentals

The most interesting point, however, is that all of this is happening before the fundamentals show any evident deterioration. An anticipatory revaluation is already under way. The market is not waiting to see the numbers. It is already moving against those that will lose value in a world where the cost of producing information, services and knowledge tends to zero.

And this is where the situation becomes truly precarious. Because this transformation is not unfolding within a stable system. It is occurring in a context in which markets are already under pressure from other factors: a higher cost of capital, uncertain growth, elevated expectations, fragility in private credit and additional sources of stress. AI is not arriving in a solid world. It is arriving in a world that is already stretched. And when a force of this magnitude enters a fragile system, the outcome is rarely linear.

The truth, however uncomfortable, is that artificial intelligence will not destroy the world of work in the simplest sense of the term: it will transform it, creating new opportunities as every major technological revolution has done. What it will destroy, however, is something far more significant: business models built on scarcity. And when that happens, the market does not merely adapt gradually. It reacts in advance — often violently — redistributing value between those capable of integrating the technology and those who are, instead, overwhelmed by it.

The question that matters

The question every investor should begin asking is not whether AI will continue to grow. That much is already clear. The real question is a different one: which companies are today selling something that tomorrow could be produced almost for free?

Because that is where the real game will be played.

And history, from Christensen onwards, teaches one thing very clearly: when a technology begins to drastically reduce costs and broaden access, it is not the strongest that survive. It is those who are quickest to adapt.