The future does not justify any price

What the SpaceX crash teaches us about artificial intelligence, technology bubbles and the world we are building.
There is one sentence every investor, entrepreneur and innovator should keep in mind: being right about the future does not necessarily mean being right about the price.
SpaceX may well be one of the most important companies of our time. It may transform access to space, build the first global satellite connectivity infrastructure, turn Starship into the logistics backbone of a future orbital economy and even contribute to a new distributed architecture for artificial intelligence.
All of this may be true.
And yet its stock can still collapse.
The article published by Wall Street Italia on 23 June 2026 described the first major post-IPO correction: from the peak of $201.80 reached on 16 June, the stock had already fallen to $154.60. That was not the end of the descent. On 27 July SpaceX trades around $111, roughly 45% below that high and almost 18% below the placement price of about $135.
This is not necessarily the failure of SpaceX. It is the failure, at least temporarily, of the idea that any price is acceptable in order to own a piece of the future.
The market is not questioning Mars
Two concepts tend to blur during phases of enthusiasm: the industrial value of a company and the financial value of its shares.
SpaceX is a real company, with technologies that are hard to replicate, enormous barriers to entry and a proven ability to execute. In 2025 it generated consolidated revenues of $18.7 billion. The Connectivity segment, driven by Starlink, alone produced $11.4 billion in revenue and $4.4 billion in operating profit.
Yet the group still closed 2025 with a net loss of $4.9 billion, followed by a $4.3 billion loss in the first quarter of 2026. Over the same period, billion-dollar investments continued in Starship, in the Starlink constellation and in the integration of xAI's activities.
The problem is not that SpaceX lacks an economic business. The problem is that in the days after the listing the market seemed to price in, simultaneously, the future value of: the world leader in space launch; the monopolist of satellite connectivity; the primary orbital logistics infrastructure; a major military and government platform; an energy and computing operator in space; a global protagonist of artificial intelligence.
The valuation did not incorporate a single success. It incorporated the possibility that SpaceX would win almost every game at once.
When a stock comes to represent not what a company is, but everything it might become, even a small doubt produces a large correction.
The market is not necessarily saying humanity will not reach Mars. It is asking who will pay for the road, how much it will cost to build and how many years it will take before it generates an adequate return.
A great technology can be a terrible investment
Economic history is full of revolutionary technologies that destroyed the capital of their earliest investors.
Railways permanently changed trade, labour, the geography of cities and the very perception of distance. But during the British Railway Mania of the nineteenth century too many lines were financed, often on unrealistic forecasts. Costs were underestimated, capacity proved excessive, and the railway share index lost roughly 65% from its 1845 peak. The railways won. Many shareholders lost.
The same happened with the internet.
In the late 1990s it was correct to argue that the network would change every industry. It was wrong to conclude that every company with ".com" in its name would produce profits, or that any valuation was justifiable.
The telecoms boom led to spectacular growth in investment, IPOs and valuations. Then came overcapacity, collapsing spending, bankruptcies and restructurings. Most of those companies disappeared, but the fibre networks built during the euphoria later helped make the digital economy possible.
This is the central contradiction of technology bubbles: they can be financially destructive and historically productive at the same time.
A bubble selects winners badly, wastes capital and pulls forward revenues that may only arrive many years later. But it also finances infrastructure, research, skills and productive capacity that a more prudent market would never have built at the same speed.
The future is often born of irrational investment. The problem is that it does not always belong to those who financed it first.
SpaceX has become three overlapping companies
To truly understand the correction, SpaceX must be seen not as a single business but as three overlapping companies.

The first is Starlink, the most concrete business: millions of users, recurring revenues, public and private customers and an operating segment that is already profitable.
The second is Starship, a gigantic industrial bet. If it succeeds, it could radically reduce the cost of reaching orbit and make economically viable activities that today still belong to science fiction. But it requires years of research, infrastructure, testing, failure and capital.
The third is the integration with xAI and orbital computing. In its SEC filings, SpaceX states that it expects to begin deploying satellites dedicated to AI computing as early as 2028. It is a fascinating vision: using solar energy, space infrastructure and satellite networks to build a new distributed computational capacity. But the same corporate documentation acknowledges the enormous capital intensity and the uncertainty surrounding these initiatives.
The investor is therefore not simply buying an aerospace company. They are financing the transition from a launch business to an integrated platform of space, telecommunications and artificial intelligence.
It is a transformation that could produce enormous value. But the more a company accumulates promises belonging to different markets, the harder it becomes to understand which part of the valuation is supported by results and which by narrative.
The most important lesson for artificial intelligence
The SpaceX crash is above all a warning for today's AI economy.
Artificial intelligence is not necessarily a fad. It is probably a general purpose technology, comparable in diffusion to electricity, the combustion engine or computing itself. But economic research shows that technologies of this kind normally require long periods of adaptation, complementary investment, new skills and a complete reorganisation of processes before delivering their full productivity effects.
The market, by contrast, tends to behave as if the transformation were immediate.
The large technology companies are investing hundreds of billions in data centres, chips, energy, networks and models. According to a Reuters analysis, by 2027 the growth in investment by the main hyperscalers could significantly exceed the growth of their operating cash flows. For every additional dollar of cash, roughly $1.57 would be invested in new capacity.
This does not prove AI is a bubble. It does show that the market is financing today a level of capacity based on tomorrow's demand, business models and margins.
That is exactly what happened with railways and telecoms.
There is, however, a fundamental difference. Tracks and fibre can retain value for decades. AI chips become obsolete far faster. A data centre designed today could require radical upgrades within a few years. Computational capacity is not immobile infrastructure: it is a permanent race against obsolescence.
That is why the decisive question is not only how much AI infrastructure we are building. It is how quickly that capacity can turn into sustainable revenue before becoming technologically obsolete.
The risk of financialising the future
Every industrial revolution passes through a phase in which financial capital runs ahead of productive capital.
At first money finances experiments. Then it finances expectations. Finally it begins to finance further expectations, ever more distant from present economic results.
That is the moment when technology stops being merely a tool and becomes a financial narrative.
With SpaceX, the market did not only buy rockets, satellites or Starlink subscriptions. It bought a symbolic stake in the future of humanity: Mars, orbital artificial intelligence, a new space economy, energy and technological independence.
When a share takes on an almost ideological meaning, valuation discipline is perceived as a lack of vision. Anyone asking when the profits will arrive is accused of failing to grasp the scale of the revolution.
But the market's task is not to decide whether a vision is fascinating. It is to determine what return the capital required to achieve it can produce.
SpaceX also raised roughly $85.7 billion through its IPO and subsequently placed $25 billion in bonds, partly to refinance bridge loans and support its development plan. It is an extraordinary financial position, but also a demonstration of how much capital is needed to turn vision into infrastructure.
Money can buy time, skills and productive capacity. It cannot eliminate technological risk, nor artificially compress the timelines of history.
Bubbles do not mean the future is false
The greatest mistake would be to read the collapse of the SpaceX share price as proof that the space economy, Starlink or artificial intelligence are illusions.

The internet bubble did not prove the internet was useless. It proved the market could not yet distinguish precisely between infrastructure, applications, imitations and real winners.
The collapse of the railway companies did not stop the railways. It separated the economically sustainable lines from those built merely to capture capital.
The same selection could happen in artificial intelligence.
Many companies today described as "AI companies" own no proprietary technology, no exclusive data, no distribution and no real competitive advantage. Some are simply adding third-party models to existing products. Others are building infrastructure that will require demand vastly greater than today's to generate acceptable returns.
AI will probably continue to spread even after a future correction. But the protagonists of the adoption phase may not be the same as those of the speculative phase.
The final winner may not be whoever builds the largest model, but whoever integrates it best into real processes.
It may not be whoever owns the most GPUs, but whoever turns computational capacity into better healthcare, more efficient industry, faster research, more accessible public services and smarter business decisions.
Technology creates value only when it stops being a demonstration and becomes an invisible infrastructure of everyday life.
What an investor should be watching today
The real question is not whether to believe in SpaceX. It is to understand how much of the future is already in the price.
One should watch Starlink's ability to keep growing without destroying margins, the real cost of the Starship programme, the amount of capital absorbed by AI, the speed at which investment converts into cash flow, and the quality of governance.
Above all, one should distinguish optionality from certainty.
Starship may open new markets. Orbital computing may become strategic. Artificial intelligence may transform the world's productive system. But an opportunity is not yet revenue, revenue is not yet profit, and a future profit does not hold the same value as a present one.
It is precisely this sequence that enthusiasm tends to erase.
The future is intact. It is the price that has changed
The SpaceX crash is not necessarily a verdict on the quality of the company. It is a referendum on the speed with which the market decided to convert distant possibilities into present value.
It teaches us that great technological revolutions do not advance in a straight line. They move through euphoria, overinvestment, error, failure, consolidation and finally productive application.
It teaches us that bubbles are not always built around something false. They are often built around something profoundly true, but anticipated too early, financed badly or valued without limits.
And it teaches us a distinction essential to understanding the age of artificial intelligence: the future may arrive exactly as it was imagined, but not necessarily in the timing, the manner or through the companies the market had bet on.
SpaceX may go on to build the most important space infrastructure of the twenty-first century. Artificial intelligence may become the foundational technology of the new economy.
But no great vision suspends the laws of capital.
The future has not been cancelled.
It has simply been repriced.