The Market Must Fall Before the Renaissance Can Rise. We Are Closer to the Fall Than Most Realise. Space. AI. Quantum. | Capital Street FX

On the morning of June 12, 2026, a rocket company listed on the Nasdaq stock exchange and by lunchtime was worth more than the combined GDP of Australia and the Netherlands. By the close of trading on June 16, it had gained another 40%. Rocket Lab, one of its closest competitors, fell 18% the same day it listed — not because anything went wrong with Rocket Lab’s business, but because the gravitational pull of the new arrival was so great that capital simply flew toward it, leaving everything in its orbit diminished by comparison.

The rocket company was SpaceX. The ticker was SPCX. The valuation was $2 trillion. And the price-to-sales ratio — the multiple of annual revenue at which it was trading — was 73 times. To understand what that number means: Amazon went public in 1997 at 7 times sales. Google listed in 2004 at 14 times. Both were called absurdly overvalued by the serious analysts of their day. Both went on to define the next twenty years of economic life. SpaceX listed at 73 times. Whether that is visionary or delusional is the central question of our time.

But here is what makes June 2026 genuinely different from every other technology moment in recent memory. SpaceX is not alone. In the weeks surrounding its listing, two other companies — Anthropic and OpenAI — have filed for what will almost certainly be the two largest IPOs in the history of public markets. Combined, they are valued at close to $2 trillion in private markets. Combined, they are projected to lose approximately $40 billion this year. And running quietly in the background, a collection of quantum computing companies are solving problems that classical computers cannot touch — working in near-total commercial obscurity, the way the internet worked in 1993 before anyone understood what it was going to become.

Three genuinely transformational technologies. Entering public markets simultaneously. At the exact moment the S&P 500 sits at the second-most expensive level in 145 years of recorded financial history.

This has happened before.

The numbers from the last two occasions are worth sitting with for a moment. When the CAPE ratio last approached current levels — it peaked at 44.2 in December 1999 — versus 40.43 today — the NASDAQ fell 78% over the following 30 months, from 5,048 to 1,114. It did not recover its March 2000 peak until April 2015: fifteen years. The S&P 500 dropped 49% and took until May 2007 to recover — then was immediately hit by 2008, losing another 57%. An investor who put money into the S&P 500 in January 2000 and held for a full decade ended the period with a net loss. In the 1929 episode, where the CAPE reached 32 before the crash, the Dow Jones fell 89% over 34 months and did not recover its nominal 1929 peak until 1954 — twenty-five years later. These are not tail risks. They are what actually happened the last two times the market was priced the way it is priced today.

Not with these specific technologies — but with this exact combination of ingredients. A real revolution meeting a priced-for-perfection market. History has two near-perfect parallels. The first was 1920–1929, when radio, automobiles, and commercial aviation arrived together in a market that was pricing in a permanent new era of prosperity. The second was 1997–2000, when the internet, telecommunications, and biotech arrived together in the longest bull market in American history. Both produced identical arcs: extraordinary wealth creation for those who understood the pattern, catastrophic losses for those who only understood the narrative. And both — crucially — were followed by an era of genuine dominance by the technologies that had looked like they were dying in the wreckage.

The question this article answers is not whether AI, space, and quantum computing will change the world. They will. The question is something older and more important:

Does changing the world mean the companies doing it right now are good investments right now?

History’s answer to that question is not the one most people are hoping for. But embedded inside that answer is the most powerful investment insight the historical record contains. And that is what this article is about.

Before we look at the evidence — and there is a great deal of it — it helps to have a map. Across 389 years of documented financial history, every major technology revolution has passed through five stages in the same sequence. The stages are not a theory. They are a description of a mechanism that repeats because the underlying drivers — human psychology, institutional incentives, the pace of technological development — do not change.

Understanding which stage you are in is not merely interesting. It is the most practically valuable piece of investment analysis you can do. The stage determines the right portfolio posture, the right instruments, the right time horizon, and most importantly — it determines who wins and who gets destroyed.

The most important thing to understand about this framework is what it says about the relationship between the technology and the investment. They are not the same thing. They are governed by entirely different timelines and entirely different logic.

The railways were a real revolution. They transformed Britain, then Europe, then the world. But of the hundreds of railway companies that went public between 1844 and 1847, almost all of them failed. The investors who backed them lost most or all of their capital. Cornelius Vanderbilt — the man who eventually built the dominant railway empire — did not invest during the mania. He waited for the wreckage and bought what was left at sane prices in the 1850s.

The internet was a real revolution. It created tens of trillions of dollars of economic value. But of the 446 internet companies that went public in 1999 alone, fewer than ten account for the overwhelming majority of value that exists today. The other 436 are gone. The investors who bought them — at CAPE ratios that look modest compared to today’s — lost most of what they put in.

This is where most analysis of the current moment goes wrong. It confuses two separate questions. The first: will AI, space, and quantum computing transform the world? Almost certainly yes. The second: do the current market prices reflect that transformation in a way that allows investors to profit from it? That is a very different question — and history is much less optimistic about the answer.

The most dangerous investor in any Stage III mania is not the naive amateur who doesn’t understand the technology. It is the sophisticated analyst who understands the technology perfectly and assumes — reasonably, but incorrectly — that understanding the technology is the same as understanding the investment. It is not. It never has been. The pioneer and the winner are different people in almost every case the historical record contains.

What follows is not a selection of convenient examples chosen to support a thesis. It is every major technology-driven market cycle in the documented financial record, from the Dutch tulip craze of 1634 to the present. The data varies in reliability by era — the 17th-century numbers come from reconstructed price series and historical accounts, while anything from 1900 onwards comes from primary sources. The pattern, however, is consistent across all eras where sufficient data exists to test it.

Read this table carefully. The column that matters most is the last one.

Notice the rows highlighted in amber. The 1920s and the late 1990s are the only prior episodes where three separate transformational technologies entered public markets simultaneously. In both cases, the peak was followed by the most severe declines in the dataset — 89% for the 1929 episode, 78% for the NASDAQ in 2000–2002. And in both cases, recovery measured in decades, not months.

Notice also what the Japan column says. The Nikkei peaked in December 1989. It did not return to that peak until February 2024. An investor who bought the Nikkei in December 1989 — at a market that was expensive but justifiable against Japan’s extraordinary growth story — waited thirty-five years to break even. Not three years. Not five years. Thirty-five years. Japan is not an anomaly. It is the most extreme version of the mechanism that appears in every row of this table.

The dot-com parallel deserves extra attention because it is the most recent, the most documented, and the most directly analogous to today. Let us spend a moment with the specific numbers.

In 1999, Cisco Systems was the most important networking company in the world. Its routers and switches built the physical infrastructure of the internet. The technology was absolutely real. The company’s competitive position was genuinely dominant. And the stock fell from $80 to $9 between 2000 and 2002 — a decline of 89%. Twenty years later, in 2020, the stock finally returned to its 1999 high. That is not a story about Cisco failing. That is a story about what happens when a genuinely great company is priced at 100 times sales at a market peak.

Amazon fell from $113 to $5.51 between 1999 and 2001. That is a decline of 95%. The investors who held through that decline and bought more — the ones who believed in the company and could afford to be patient — are now holding one of the greatest investments in financial history. But the investors who bought at $113 in 1999 and sold at $5.51 in 2001 are not those investors. They are the investors who funded the infrastructure Amazon would use to dominate the next twenty years. They were right about everything except the price and the timeline.

The three technologies at the centre of the current moment are not equivalent. They are at different stages of development, different stages of commercial maturity, and different stages of market pricing. Understanding where each one sits in the Five-Stage framework is the entire analytical challenge — and the honest answer is more nuanced than most coverage of this moment allows for.

Space — The Infrastructure Argument Is Real. The Price Is the Question.

SpaceX is not a speculative technology company. It is the only organisation on earth that has cracked reusable rocket economics at scale, reducing the cost per kilogram to orbit from $60,000 in the Space Shuttle era to approximately $2,700 today — a 96% cost reduction in fifteen years. Starlink is a genuinely profitable satellite internet business serving over 4 million subscribers with a structural moat that no competitor can approach without a decade and tens of billions in investment. These are real assets generating real cash flows.

At 73 times sales, however, the stock price is not asking whether SpaceX is a good business. It is asking whether SpaceX will become the dominant infrastructure company of the next century, commercialise orbital manufacturing, establish a viable Mars economy, and maintain that dominance against well-capitalised competitors for decades to come. All of those things may happen. None of them are priced with any margin of safety at $174.90 per share.

The smaller space companies — Rocket Lab, AST SpaceMobile, Intuitive Machines — sit at more interesting valuations but carry the risk of direct competition from the company that just raised $75 billion and can afford to price aggressively in any market they choose to enter.

Artificial Intelligence — The Most Important Technology in a Generation. The Most Expensive in History.

Artificial intelligence is a General Purpose Technology — the classification economists give to technologies that diffuse across every sector of the economy and generate complementary innovations for decades. The steam engine was one. Electricity was one. The internal combustion engine was one. The evidence that AI fits this category is increasingly strong: measurable productivity gains of 20–40% for knowledge workers using frontier models, enterprise deployments at scale across medicine, law, finance, and engineering, and genuine disruption of workflows that have been unchanged for a generation.

But consider the specific numbers on the table above. Anthropic is valued at $965 billion in private markets. Its revenue run rate is $47 billion annualised. That gives it a price-to-sales ratio of approximately 20 times — high but not insane for a high-growth enterprise software company approaching its first profit. OpenAI, by contrast, is valued at $852 billion with $25 billion in annualised revenue — approximately 34 times sales — and is projected to lose between $25 and $27 billion this year on a GAAP basis. It has committed $600 billion to infrastructure spending and does not expect to generate positive cash flow until 2030. The implicit comparable for OpenAI’s valuation is Microsoft — the operating system on which the next generation of businesses is built. The financials it will disclose look considerably more like Google.

Both companies are real. Both are building something genuinely important. The question — as it was with WorldCom, as it was with Cisco, as it was with every great infrastructure builder at the peak of Stage III — is whether the price reflects the business as it exists in 2026 or the business as it might exist in 2036.

Quantum Computing — The Transistor Moment. The Most Patient Opportunity.

Of the three technologies, quantum computing is the one where honesty demands the most patience and the most humility. In December 1947, three physicists at Bell Labs demonstrated the transistor — a device that would eventually underpin every piece of electronic technology on earth. The first commercial transistor radio appeared in 1954. The integrated circuit arrived in 1958. The personal computer did not appear until 1977. Intel, which would dominate the semiconductor industry for forty years, was founded in 1968 — twenty-one years after the transistor’s invention.

In 2024, Google’s 105-qubit Willow chip achieved what the quantum computing field had been working toward for two decades: below-threshold quantum error correction, meaning that adding more physical qubits actually reduces error rates rather than amplifying them. This is the equivalent of the 1947 transistor demonstration. The physics is proven. The engineering challenge is what remains — and that challenge, if the semiconductor parallel holds, will take the better part of a decade to translate into commercial applications at scale.

The market for quantum computing is projected to grow from $1.4 billion today to $4.2 billion by 2030 and $170 billion by 2040 — a 120-fold expansion. The semiconductor market grew approximately 150-fold in the sixteen years following the integrated circuit. The quantum projections are consistent with the historical parallel. But the timeline implies that the real commercial payoff is a decade away at minimum. The investors who buy quantum companies at speculative valuations today are pricing in a world that may not arrive until 2035. The investors who will make the most money from quantum are probably the ones who will buy in Phase IV — when the technology is real and proven and the market has forgotten it existed.

There is a question worth asking before we look at the valuation data. Why does it matter that the broad stock market is expensive when the specific technologies driving the current excitement are real and transformative? The answer is the most important structural point in this entire analysis.

When the broad market is priced at extreme valuations, it is because capital has been flowing broadly into equities — including the technology sector — on the assumption that the future will be better than the past. When that assumption cracks, it does not crack selectively. The entire asset class reprices simultaneously. The revolutionary companies fall with the mediocre ones. Amazon fell 95% in the dot-com crash — not because Amazon was a bad company, but because it was a technology company in a technology sell-off in which the price of capital rose dramatically and the tolerance for losses evaporated overnight.

The broad market valuation is not a separate problem from the technology sector valuation. It is the same problem. And the current readings on the three most reliable long-run measures we have are among the most extreme in recorded history.

These numbers do not tell us when the correction arrives. They never do. The dot-com CAPE of 44 was flashing extreme readings for three years before the actual peak. Japan’s valuation signals were extreme for years before December 1989. The mechanism is not a countdown timer. It is a measure of how much margin for disappointment exists in the current price — and at 40.43, the answer is very little.

What the numbers do tell us with reasonable confidence is the shape of the distribution of likely outcomes. At current valuations, the probability of the S&P 500 delivering its historical median return over the next decade is very low. The probability of a significant correction at some point in the next 1–5 years is, by historical precedent, very high. And the probability that the technology companies currently commanding the highest valuations will be the specific companies that dominate in 2040 is — based on every prior cycle in the table above — considerably lower than their current market prices imply.

This is not a prediction. Markets are not predictable on any precise timeline, and anyone who claims otherwise is selling something. What follows is a probability-weighted map derived from the historical record — the same record we have just examined in detail. Think of it as the shape of what tends to happen, not a guaranteed schedule of events.

History provides a very specific profile of the companies that reach Stage V dominance. They are almost never the most celebrated names at the Stage III peak. They share a structural characteristic: their economics work even if you strip away the technology label. If you looked at them purely as businesses — ignoring the AI story, the space story, the quantum story — you would still want to own them. The technology narrative is upside. The business is the foundation.

We have organised every meaningful company in this space into four tiers. The tiers are not a judgment of the technology — they are a judgment of the survivability of the business model through Stage IV and the probability of Stage V dominance.

Every trade below flows from the same argument the article has made. The market is at a historic valuation extreme. A correction is more probable than most investors currently believe. That correction, when it arrives, will precede a period of genuine technological prosperity — the Renaissance the title describes. The trades are positioned for both sides of that arc: short the peak, accumulate in the correction, hold for the era that follows.

These are educational trade frameworks, not personalised investment advice. All levels are based on publicly available market data as of June 20, 2026. All investments carry risk of loss. Seek independent professional advice before trading.

Let us return to that morning.

June 12, 2026. A rocket company lists on the stock exchange. By lunchtime it is worth $2 trillion. In the weeks before and after, two AI companies file for IPOs that will collectively raise more money than any technology companies in history. Quantum computers are solving problems in minutes that would take classical computers longer than the age of the universe. The S&P 500 is priced at the second-most expensive level in 145 years of recorded data.

None of this is an illusion. The rockets are real. The AI is genuinely changing how people work, think, and create. The quantum computers are solving real problems. The transformation that is being priced into today’s markets will, in some form and on some timeline, materialise.

The question has never been whether the transformation is real. Every transformation in the historical record — the railways, the automobile, the electricity grid, the internet — was real. The question has always been the same question, in every era, without exception: does the fact that the transformation is real mean that the current price is a good price to pay for exposure to it?

And the answer — drawn from 389 years of evidence, from every crash-and-recovery cycle in the documented record — is: sometimes yes. Sometimes no. And the difference between those two outcomes is not determined by the technology. It is determined entirely by the price at which you buy it.

George Hudson was not wrong about the railways. He was one of the most important figures in the construction of the British rail network. He was wrong about the price at which his companies were capitalised, the leverage they carried, and the timeline over which the revenues would materialise. The investors who backed him paid for that error with everything they had. The investors who came after him — who bought the surviving railways at prices that made economic sense, who consolidated the wreckage into viable businesses — built the infrastructure of Victorian Britain and the fortunes that came with it.

The investors who bought Amazon in 1997 at 7× sales had conviction in a real transformation and paid a price that, even in retrospect, was not insane. The investors who bought at the 1999 peak at 150× sales had identical conviction in the identical transformation and paid a price that destroyed most of them, even though the company went on to be one of the most valuable in history.

We are not saying the current moment is 1999. We are saying the valuation signals are at approximately 1999 levels. We are saying the same three-technologies-arriving-simultaneously pattern that characterised 1920–1929 and 1997–2000 is present now. We are saying that every prior episode of this specific configuration ended in a Stage IV correction of significant magnitude. And we are saying that after every one of those corrections — every single one — the technologies continued to develop, the best companies continued to grow, and the investors who understood the pattern well enough to hold cash, buy in the wreckage, and stay patient through the Valley came out the other side in possession of something that looks, from a sufficient distance, like wisdom.