The Price of Belief: Why Every Boom Ends the Same Way

Jun 10, 2026

10 min read

3 views

87%

The strange thing about bubbles is that they always feel justified

What if the most dangerous sign in a market is not euphoria, but competence?

That sounds backwards, because bubbles are usually described as irrational episodes, moments when people throw caution aside and buy nonsense. But the deeper pattern is more unsettling: bubbles are rarely built on pure fantasy. They are built on a real technological shift, a believable future, and just enough evidence to make the impossible seem inevitable. That is why they survive so long, attract so much capital, and only look ridiculous in retrospect.

The market today sits inside that exact tension. Public stocks are still expensive by the standards of the last decade, even after drawdowns and tightening financial conditions. Meanwhile, private and public investors are pouring extraordinary sums into artificial intelligence, as if the next great monopoly is already visible on the horizon. These are not separate stories. They are two views of the same phenomenon: a market that has not decided whether it is pricing earnings, or pricing faith.

And that distinction matters, because when capital stops pricing cash flows and starts pricing destiny, the game changes. The winners are no longer just the businesses that earn the most. They are the stories that can command the most belief.

The market is not asking what companies earn, but what future they are allowed to imagine

Traditional valuation works on a simple bargain. Investors supply capital now in exchange for a share of future profits. The price matters because it anchors the amount of future success required to justify the present. If a company trades at a modest multiple, it can disappoint a little and still survive. If it trades at an extreme multiple, the future has to arrive not just successfully, but on schedule, at scale, and with minimal friction.

That is why markets become fragile when prices remain elevated while financing conditions tighten. Rising rates do not just make money more expensive. They compress the distance between story and proof. A company can no longer rely on endless cheap capital to keep its narrative alive. The burden shifts from imagination to arithmetic.

But periods of technological upheaval complicate this logic. A new platform can create the impression that old valuation rules no longer apply. Search, mobile, cloud, and now AI all create genuine discontinuities. The problem is that investors often confuse a real discontinuity in technology with a discontinuity in economics. The first may be real. The second is much rarer.

This is why bubbles are so persuasive. They begin with an insight that is correct enough to be dangerous. The internet really did change distribution. The cloud really did change infrastructure economics. AI really may change knowledge work. Once that is true, almost any business plan can be smuggled into the future with a straight face.

A technology revolution does not eliminate valuation discipline. It raises the stakes of getting valuation wrong.

The deeper question, then, is not whether the technology matters. It is whether investors are overpaying for the right future, or financing a future that only exists in the language of inevitability.

The real bubble is not in valuation, but in compression of time

One of the most revealing features of contemporary manias is not just their scale, but their speed. Older bubbles had time to mature through visible excess. Today, capital can move from thesis to frenzy in a fraction of that time. The result is a peculiar kind of acceleration: not simply more money, but more money arriving before institutional memory can update.

That is why comparisons to past episodes matter. The internet bubble did not merely involve high valuations. It involved the belief that companies should “get big fast,” that distribution advantages would outrun economics, and that infrastructure spending could be rational even if profits were distant. Some of that was true. Much of it was not. The result was a wave of startups and public companies that mistook a network effect for a business model.

Artificial intelligence is inheriting the same dynamic, but with a sharper edge. The leading firms are not only raising enormous sums, they are committing to spending patterns that assume a vast future market. In practice, that means the industry is behaving as if scale itself is a moat, regardless of whether the eventual economics support it.

This creates a crucial distortion: capital expenditure becomes a proxy for conviction. The bigger the spend, the more credible the story appears. But the bigger the spend, the more catastrophic the eventual correction if the story disappoints. In a sense, the market is financing not just product development, but a public test of belief.

That is the hidden danger of speed. A slower bubble at least gives reality time to intervene. A fast bubble can build so much momentum that many stakeholders are committed before the first serious doubts appear. By the time the model begins to wobble, the system has already normalized the assumptions that made it wobbly.

A useful analogy is a bridge built from confidence rather than steel. Each new investor crossing it makes the bridge seem stronger, because traffic is visible. But traffic is not structural integrity. The market often mistakes use for durability.

Why the most dangerous company is not the obvious fraud, but the plausible monopoly

The most unnerving part of today’s boom is not that some companies may fail. It is that a few may succeed so spectacularly that they justify the excess around them. That is what makes modern bubbles hard to detect in real time. They are not built entirely on nonsense. They are built on asymmetry.

If one company in a generation becomes the operating system of a new technological layer, it can change the economics of an entire sector. Search did that. Smartphones did that. Cloud did that. So when investors see a new platform taking shape, they do not just price the company in front of them. They price the possibility of winning the layer beneath everything else.

That is a rational impulse, but it becomes dangerous when applied too broadly. A world in which every AI company is the next platform is not a world with one platform. It is a world with too many claims on a future that cannot support them all.

The crucial distinction is between category creation and category capture. Many firms can help create a category. Very few can capture it. During the bubble phase, investors often pay as if creation automatically implies capture. That is why the valuations look sane in the language of market share, but insane in the language of probability.

A second distinction matters just as much: capability versus control. A company may have brilliant researchers, a magnetic founder, and a compelling prototype. Yet the long-term prize in technology is rarely just capability. It is control over standards, distribution, data, and customer dependency. Without those, technical excellence can become expensive differentiation rather than durable power.

This is where the comparison to historic monopolies becomes illuminating. The enduring giants did not merely build products. They built chokepoints. They sat at layers of the stack where the rest of the economy had no choice but to route through them. Investors in every cycle are tempted to buy the dream of such a choke point before anyone knows whether it exists.

And that is the real peril. A plausible monopoly attracts capital like gravity, even when the evidence is incomplete.

A better way to think about bubbles: they are tests of narrative leverage

Most people think bubbles are about greed. That is too shallow. Greed is constant. What changes is the amount of narrative leverage available in the market.

Narrative leverage is the ratio between the quality of a story and the amount of capital it can attract before the story is verified. When narrative leverage is low, markets demand evidence early. When it is high, they finance the story far ahead of proof.

Technology cycles increase narrative leverage because they create credible ambiguity. Nobody knows exactly how the new world will work, so almost any well-crafted vision can appear reasonable. Investors are not buying certainty. They are buying optionality. But optionality is frequently mispriced because it is emotionally easier to believe in upside than to model downside.

This is especially true when there is a precedent that seems to forgive excess. Amazon is the classic example. It looked absurd for years, and in some respects the criticism was right. The eventual outcome was extraordinary, but it required a second act that few people anticipated. The lesson investors often take from that story is that overinvestment in the right company is fine. The more accurate lesson is that a rare outcome can vindicate a bad price only after a long and dangerous delay.

That delay is what makes bubbles socially seductive. They create the sense that skepticism was premature, which then trains a new generation to distrust caution. The problem is not that skepticism was wrong in principle. It is that the market’s time horizon was longer than the skeptics expected.

This is why bubbles and breakthroughs are entwined. Every genuine technological platform generates a population of mistaken skeptics and a larger population of mistaken believers. The hard task is not to choose cynicism or enthusiasm. It is to distinguish between the kind of uncertainty that rewards patience and the kind that simply subsidizes overconfidence.

In a bubble, the market is often right about the existence of a revolution and wrong about who will profit from it.

That sentence is uncomfortable because it suggests that both sides can be half right, and that is often the most dangerous state of all.

The practical lesson: price the path, not the promise

If there is a usable framework here, it is this: do not ask whether a company belongs to the future. Ask what must happen, in sequence, for the price you are paying to become defensible.

That means evaluating not only the destination, but the road.

For any story stock or venture bet, break the thesis into four questions:

What has to be true technically? Does the product actually work at the level required for mass adoption, and does it improve fast enough to stay ahead of alternatives?
What has to be true economically? Can the company acquire customers, retain them, and serve them at margins that justify the capital intensity?
What has to be true competitively? Does the firm own a durable choke point, or is it simply early?
What has to be true financially? How much capital is required before the business becomes self-sustaining, and what happens if the cost of capital rises before then?

This approach is more useful than asking, “Is this the next big thing?” because it forces you to price the transition, not the mythology. It also reveals why many boom era businesses look compelling in slides but fragile in reality. Their path to greatness depends on too many unproven miracles arriving in the right order.

The market’s job is not to reward the most exciting future. It is to determine how much the present should pay for access to that future. Those are not the same question. In fact, confusing them is often the first sign of excess.

Key Takeaways

Do not confuse technological truth with investment truth. A real innovation can still be a bad buy if the price assumes perfection.
Watch the speed of capital, not just the amount. Fast inflows compress judgment and make mistakes compound before they are visible.
Separate category creation from category capture. Many firms help build a new market, few control it.
Price the path to profitability, not the promise of transformation. Ask what must happen in sequence for the valuation to make sense.
Treat huge numbers as a warning signal, not a victory lap. When spending becomes too large to intuit, discipline should increase, not decrease.

The market eventually stops paying for possibility and starts paying for proof

Every boom tells the same story in a different costume. First comes a genuine shift, then a compelling narrative, then capital that arrives faster than caution can organize itself. Eventually, arithmetic returns. Not because the technology was fake, but because the market ran ahead of the economics.

That is the lesson embedded in both valuation and mania. High prices do not merely predict disappointment. They change behavior. They loosen standards, accelerate spending, and turn uncertainty into a competitive pressure to participate. Once that happens, the market is no longer evaluating the future. It is helping create a future that may or may not be affordable.

The hardest insight is also the most useful: the purpose of valuation is not to kill ambition, but to discriminate between ambition that compounds and ambition that incinerates capital.

So the next time a company, sector, or index looks expensive, do not ask only whether it could be right. Ask what kind of world must exist for it to be right, who captures the value in that world, and how long the bridge to that world has to hold.

Because in every great bubble, the market is not merely buying shares. It is buying the right to believe that this time, the numbers can be postponed.

And eventually, they cannot.