Reading time: 8 minutes
Between the promise of technological revolution and the sudden speculation, the border becomes thinner and thinner.
While investment announcements are in tens of billions of dollars and every technology giant claims "revolution IA", a question must be asked: Are we not witnessing the formation of a new financial bubble?
Between colossal infrastructures, promises of productivity and economic models that are still weak, artificial intelligence today attracts capital flows comparable to those of the Internet bubble of the 2000s.
But behind euphoria lies a more contrasting reality: the technological break is undeniable, profitability remains hypothetical.
Let us make a lucid reading of this global rush towards AI, where market enthusiasm sometimes seems to be ahead of economic reason.
Understanding the mechanisms of a financial bubble: lessons from modern history
Financial bubbles are not isolated accidents: they are the recurring manifestation of over-optimism.
They are always born with the same responsibility: innovation, deregulation or a new belief triggers a rush of investors convinced to enter an era of unlimited prosperity.
Universal springs of bubbles
A bubble follows a mechanism now well identified:
- A catalytic innovation (technology, financial product, new market) triggers interest and attracts capital.
- Exercise effect amplify the movement: investors and institutions follow by mimicry those who seem to succeed.
- The disconnection of fundamentals : valuation feeds on anticipations more than real income.
- Liquidity illusion Everyone thinks they can get out in time.
- The turnaround : an exogenous shock, an increase in rates or a simple reversal of psychology causes the collapse.
Japan of the 1980s: the real estate bubble and the "lost decade"
In the years 1985-1991, Japan experienced unprecedented economic euphoria. Very low interest rates and financial deregulation trigger frantic speculation about real estate and equities.
In Tokyo, the square metre reaches surrealistic values; Some of the land in Ginza is worth more than all of California.
Banks lend indiscriminately, businesses use their land assets as leverage, and the Bank of Japan is slow to intervene.
When rates rise, the bubble bursts:
- Real estate prices fall by more than 60%;
- The Nikkei loses more than half its value;
- Japan enters into a prolonged economic stagnation, marked by the deflation and weakening of the banking system.
Lesson An excess of liquidity and confidence can ruin an even industrialized and disciplined economy.
1995-2000: the Internet bubble and "rational exuberance"
With the explosion of the Internet in the late 1990s, a new technological promise emerged: that of a connected world, without borders or intermediaries.
Thousands of start-ups ".com" are born; Many have neither sales nor products nor customers.
The financial markets are burning up: the Nasdaq triples in value in two years, driven by the belief that a "new economy" will make the classic laws of profitability obsolete.
In March 2000, reality caught up with fiction: the bubble broke out, valuations collapsed, and $5 trillion in capitalization disappeared.
Many companies will not survive, but from this purge will be born Amazon, Google and the foundations of the modern digital economy.
Lesson : a technology can be revolutionary without justifying unlimited valorisation.
2007-2008: the subprime crisis and the illusion of deregulated finance
The 2000s saw another type of innovation imposed: securitisation. US banks turn mortgages into financial products sold worldwide.
Real estate seems never to have to decline, households are heavily in debt, and the markets consider these securities to be safe.
When the rates rise, the defects chain, the titles devalue, and the domino effect leads to the collapse of Lehman Brothers.
The crisis is becoming global: frozen credit, widespread recession, bank bankruptcies and mass unemployment.
Lesson A financial innovation without accounting safeguards or prudential control turns the promise of return into a systemic disaster.
The same scenario that repeats itself
From Tokyo to Wall Street, the pattern is always the same:
- a seductive innovation,
- one capital inflows,
- a overvaluation of assets,
- then one brutal return to reality.
Each bubble, whether real estate, technological or financial, translates the same imbalance between hope and value created.
Today's question is: Is artificial intelligence the new avatar of this story?
Artificial intelligence: a rush to digital gold
Since 2023, artificial intelligence has become the new founding myth global growth.
Investment announcements reach record levels:
- Alphabet (Google) plans more 90 billion $ capital expenditure in 2025;
- Meta (Facebook) enhances its capex to 70 billion $ ;
- Microsoft a growing share of its cash flow is devoted to its partnership with OpenAI.
These amounts not only fund software: they are used to build a colossal industrial ecosystem Data centers, GPU, fibre optic networks, cooling systems, and dedicated energy sources.
The International Energy Agency estimates that the electricity consumption of data centres could double by 2030, the AI represents more than half of this growth.
We are witnessing a planetary race to computing power, where each country wants to secure its access to "digital gold".
Profitability still very low
Despite this effervescence, income from AI remains embryonic.
- IA products (ChatGPT, Copilot, Gemini, Claude, etc.) represent less than 5 % the turnover of their parent homes.
- Operating costs are staggering: the cost of an AI request sometimes exceeds 500 times that of a classical search request.
- Most specialized start-ups do not yet generate profits and depend on fundraising.
In other words, capital still precedes economic models.
This asymmetry (immediate expenses, deferred revenues) is the classic fuel of a speculative bubble.
Current symptoms of a bubble in formation
| Symptoms | Current event |
|---|---|
| Extreme stock market concentration | The « Magnificent Seven » represent over 35% of S&P 500higher than 2000. |
| Infrastructure development | Titles related to GPUs, data centers or cooling exchange to multiples > 50 x. |
| Effect of mimicry | All companies claim an "IA plan" to seduce investors. |
| Energy voltages | Power supply constraints become the main obstacle to growth. |
| Unbalanced economy | US growth is artificially supported by AI technology enhancements. |
These signals recall the excesses of the years 1999–2000: Anticipation of earnings far exceeds their materialization.
What differentiates (again) the IA bubble from the previous
However, the comparison should be nuanced.
Current leaders (Microsoft, Alphabet, NVIDIA, Amazon) have solid cash, real profits and tangible assets.
Data centres and energy infrastructure are Reusable for other activities (cloud, cybersecurity, hosting). Their use value will therefore not collapse to zero, even in the event of a slowdown.
Another major difference is that energy and material constraints play a role here. natural brake role.
The electrical capacity, the land available and the manufacturing of GPUs impose physical limits on packaging, which did not exist during the Internet bubble.
The AI operates in a "real" environment: it is a possible bubble, but backed by tangible industrial foundations.
Possible correction scenarios
Three catalysts could initiate market deflation:
- Declining profitability If AI services fail to monetize to the level of investment.
- Energy - delays in connection with connection, increased electricity costs, environmental constraints.
- Sustained increase in interest rates : more severe updating of future cash flows and contraction of valuations.
A "soft landing" is still possible, but a sudden slowing of the capex or a collapse of a major actor could act as a trigger.
A bubble of anticipation, not pure speculation
Artificial intelligence is not a bubble without substance.
It is the vehicle for a profound transformation of work, production, marketing and research.
But the market anticipated too fast the economic benefits of a technology to be integrated years.
What we are experiencing is not an empty bubble, but an empty bubble. Anticipation bubble : the gap between the potential for disruption and its actual profitability.
It will not necessarily collapse as in 2000, but will slowly deflate As actors distinguish the useful tool from the expensive gadget.
The economic reason for technological enthusiasm
The financial history shows that each major innovation (railway, electricity, Internet) has first generated a speculative euphoria phase.
Artificial intelligence will not escape. His promise is immense; its profitability, still uncertain.
Innovation is never immune to speculation.
The cycles of euphoria and disillusionment are the natural breath of capitalism.
For investors, auditors and managers, caution now requires tangible indicators : cash flows, return on investment, energy efficiency, data center occupancy rates.
Because beyond the technological glamour, the true value of the AI will be measured not at the speed of the advertisements, but at the profitability of the uses.
