From Tulip Mania to Artificial Intelligence: Why Every Generation Believes It Has Reinvented Capitalism—and Why I Expect the AI Investment Boom to Face Its Reckoning by 2030
By John S. Morlu II, CPA
October 2026 | Economic History, Technology, Capital Markets & Business Strategy
The Great Illusion of Every Generation
Every generation believes it has discovered something so revolutionary that the old rules of economics no longer apply.
In the 1630s, it was tulips.
In the 1840s, it was railways.
In the late 1990s, it was the internet.
Then came digital transformation, cloud computing, the Internet of Things, blockchain, big data, machine learning, and now artificial intelligence.
Each wave arrived with remarkable promises. Industries would disappear. Productivity would explode. Fortunes would be created. Established companies would become irrelevant. Anyone unwilling to embrace the new economy risked being left behind.
And investors were repeatedly assured that this time was different.
Sometimes, it was different.
The technologies changed. Economies advanced. Entire industries emerged. But the relationship between technological possibility and financial expectations remained remarkably familiar.
The technology often delivered. The investment expectations frequently did not.
That distinction is essential.
Railways transformed transportation even as railway investors suffered. The internet transformed commerce after thousands of internet companies disappeared. Blockchain introduced new ways of recording and transferring digital assets while speculative projects collapsed. Machine learning became embedded in everyday business operations even as much of the surrounding promotional language faded.
History suggests that technological revolutions and speculative excess can exist simultaneously.
Indeed, some of the most important innovations attract the greatest exaggeration precisely because their potential is so extraordinary.
Today, artificial intelligence represents another extraordinary technological advance. Its capabilities are real, its commercial applications are expanding, and its long-term economic impact could be profound.
Yet I believe that before the end of 2030, the AI investment economy will experience a major shakeout—marked by failed companies, significant consolidation, valuation corrections, and a redirection of venture capital toward more defensible opportunities.
That is a forecast, not an established fact.
My argument is not that artificial intelligence will fail.
It is that the market may be overestimating how many AI businesses can become profitable, how quickly they can do so, and how much investors should pay today for earnings that may arrive years from now.
The difference between those questions may determine who builds lasting wealth and who merely finances the next generation’s infrastructure.
1. Tulip Mania, 1636–1637: When Scarcity Became a Business Model
Nearly four centuries before artificial intelligence became the center of global investment, Dutch merchants discovered that flowers could attract extraordinary prices.
Tulips had become desirable status symbols in the Netherlands. Certain varieties were rare, difficult to reproduce quickly, and visually distinctive.
Their scarcity created genuine economic value.
Speculation then carried expectations beyond ordinary horticultural commerce.
Traders began negotiating contracts for bulbs that were still underground. A buyer could acquire a claim with the expectation of selling it to someone else at a higher price.
The market was no longer concerned only with the beauty of a flower. Increasingly, it was concerned with the willingness of the next buyer to pay more.
Prices for prized varieties rose dramatically before the speculative market broke down in early 1637.
However, the popular version of Tulip Mania deserves an important correction.
Historian Anne Goldgar’s archival research challenges the familiar stories of widespread national bankruptcy and ordinary Dutch citizens collectively losing everything. The speculative trade involved a relatively limited network, and its wider economic damage appears to have been much smaller than popular accounts suggest.
That distinction matters. A serious examination of bubbles should not begin by repeating exaggerated stories about an earlier bubble.
The lesson is not that Dutch society collectively lost its senses. It is that a legitimate market for a scarce product developed expectations that proved unsustainable.
Tulips did not disappear after 1637. The Dutch flower industry endured.
The speculative prices did not.
Four centuries later, the Netherlands remains associated with a globally significant flower trade.
The real economic activity outlived the mania.
Research: Anne Goldgar, Tulipmania (2007); Smithsonian Institution; Rijksmuseum.
2. Railway Mania, 1840s: The Infrastructure Was Real. So Were the Losses.
Railway Mania offers an even more relevant comparison with artificial intelligence because railways required enormous amounts of physical infrastructure.
During the 1840s, Britain experienced a surge in railway investment. Investors correctly understood that rail transport could transform manufacturing, commerce, geographic mobility, and national economic integration.
They were right about the technology.
They were not always right about the investments.
Railway companies and promoters raised capital for new routes. Some schemes were commercially questionable, some were overextended, and others relied on optimistic assumptions about future traffic and profitability.
The Federal Reserve Bank of New York’s historical analysis documents how railway shares multiplied in number and rose sharply in price before enthusiasm weakened in 1845. The broader commercial crisis followed in 1847, influenced by railway speculation alongside harvest failures, financial tightening, and other economic pressures.
Importantly, the subsequent railway network provided substantial economic benefits.
The tracks remained. Transportation improved. Trade expanded.
But the existence of useful infrastructure did not guarantee attractive returns to the investors who financed it.
This is perhaps the most important historical comparison for today’s AI infrastructure boom.
Data centers, advanced chips, cooling systems, electricity networks, and computing facilities may provide lasting economic benefits.
But useful infrastructure can still be overbuilt, overfinanced, or acquired at excessive prices.
A railway can change the world without making every railway shareholder wealthy.
A data center may do the same.
Research: Federal Reserve Bank of New York, Railway Mania and the Commercial Crisis of 1847; Business History research on railway investment and profitability.
3. The Dot-Com Boom, 1995–2002: The Internet Won. Thousands of Investors Did Not.
The dot-com era was one of the clearest modern demonstrations of the difference between a technological revolution and an investment bubble.
During the late 1990s, the internet was rapidly connecting businesses and consumers.
Entrepreneurs promised to replace physical stores, reshape advertising, remove intermediaries, and create entirely new commercial markets.
Many of those predictions eventually came true.
Unfortunately, investors frequently paid for that distant future as though it had already arrived.
Companies attracted large valuations despite limited revenue, uncertain margins, expensive customer acquisition, and no demonstrated path to profitability.
Some believed that attracting visitors was a substitute for earning money.
Others assumed that rapid expansion would eventually solve every weakness in their business models.
The market rewarded growth stories, particularly those supported by enthusiastic venture capital and favorable public-market conditions.
Then the financing environment changed.
The Nasdaq Composite peaked on March 10, 2000. By October 2002, it had lost approximately 78% of its value.
Companies such as Pets.com and Webvan became symbols of a broader collapse in speculative internet businesses.
But the internet did not disappear.
On the contrary, online commerce, search, cloud services, digital advertising, and internet-based communication continued to develop.
Amazon survived. Google grew into a major business. Entirely new digital industries followed.
The enormous correction did not prove that the internet was worthless.
It demonstrated that investors had often paid too much, too early, for uncertain future earnings.
The dot-com collapse was not the death of the internet. It was the end of an investment environment in which almost any business could become attractive merely by adding “.com” to its name.
That is the distinction today’s AI investors should study.
Research: J. Bradford DeLong and Konstantin Magin, A Short Note on the Size of the Dot-Com Bubble, NBER (2006); historical Nasdaq market data.
4. The Digital Economy, 2000s–2010s: When Everything Became a Transformation
After the dot-com collapse, the commercial internet matured.
Companies increasingly adopted cloud computing, smartphones, online platforms, digital payments, and connected business applications.
The phrase digital transformation became a central feature of corporate strategy.
Organizations invested in technology to improve customer experience, automate processes, reduce costs, and expand their reach.
These changes were real. Many created substantial value.
But the term also became so broad that almost any technology expenditure could be described as transformational.
A company could replace its accounting software, launch a new website, move documents into cloud storage, and announce that it was undertaking a digital transformation.
Sometimes, that description was justified.
Sometimes, the investment did little more than digitize an inefficient process.
The central managerial mistake was confusing technology adoption with organizational improvement.
A poorly designed approval process does not become efficient merely because employees now approve transactions on their phones.
A weak business model does not become strong because its reports are displayed on a colorful dashboard.
And a company does not become customer-centered simply by purchasing customer relationship management software.
Digital transformation was not one unified financial bubble equivalent to the dot-com crash. It was a broad, overlapping business transition, accompanied by numerous investment booms, exaggerated promises, and unsuccessful projects.
The lasting lesson was organizational rather than purely technological:
Technology can accelerate a good business process. It can also accelerate a bad one.
5. The Internet of Things, 2010s: Fifty Billion Devices and an Expanding Universe of Expectations
The Internet of Things, or IoT, promised to connect physical objects to digital networks.
Factory machines would predict failures. Vehicles would exchange information. Agricultural equipment would monitor environmental conditions. Buildings would manage energy consumption automatically.
The opportunities were substantial.
So were the forecasts.
In 2010 and 2011, influential industry projections popularized the expectation that approximately 50 billion devices would be connected by 2020.
The number became part of the commercial imagination.
But by 2016, several major forecasters had revised their expectations downward, as documented by IEEE Spectrum. The original forecasts also differed in what they counted as connected devices.
This illustrates a common problem in emerging-technology markets.
A forecast can be directionally reasonable while being commercially misleading.
Even if billions of devices eventually become connected, that does not mean every device requires a profitable software subscription, that every customer is willing to pay, or that every IoT startup has an enduring competitive advantage.
The enduring value of IoT has become particularly evident in areas such as industrial monitoring, logistics, energy management, and operational maintenance.
The grand narrative of connecting everything became less important than the business economics of connecting the right things.
Research: IEEE Spectrum, Popular Internet of Things Forecast of 50 Billion Devices by 2020 Is Outdated (2016).
6. Blockchain, 2017–2022: A Technology, a Token, and a Speculative Market
Blockchain offered a compelling proposition: a method of maintaining shared records without relying exclusively on a conventional central database administrator.
Applications emerged in digital assets, payments, recordkeeping, and other areas where parties needed to coordinate information or transfer value.
But commercial enthusiasm soon moved beyond the underlying technology.
Initial coin offerings, or ICOs, allowed projects to raise capital by selling digital tokens.
The United States Securities and Exchange Commission warned investors about ICO risks in July 2017, including potential fraud and misleading promises.
An SEC investor-advocate report subsequently noted that reported ICO fundraising rose from less than $100 million in 2016 to approximately $4 billion in 2017, then exceeded $16 billion during the first two quarters of 2018.
The problem was not that blockchain lacked every conceivable application.
The problem was that ownership of a token could become detached from any reasonable assessment of a project’s underlying economic value.
Investors were frequently asked to believe that a proposed technological ecosystem would create demand for a particular digital asset.
That required several assumptions to be true simultaneously: the platform would be completed, users would adopt it, transactions would grow, and the token would capture a meaningful share of the resulting value.
In many projects, those assumptions were never demonstrated.
Later failures in the broader cryptocurrency industry exposed additional weaknesses in governance, custody, financing, and institutional oversight.
The bankruptcy of FTX in November 2022 became a particularly visible example of failures in a centralized crypto business. It was not, by itself, proof that blockchain technology was fundamentally defective.
The distinction remains important.
A distributed ledger can perform a useful function without making every cryptocurrency attractive.
The existence of a technology does not establish the investment value of every asset created around it.
That principle applies just as strongly to today’s AI economy.
Research: SEC Investor Bulletin on Initial Coin Offerings (2017); SEC Office of the Investor Advocate, FY2018 report.
7. Big Data and Machine Learning, 2010s: When Prediction Became the Product
Before generative AI became a household term, businesses spent years exploring big data, predictive analytics, and machine learning.
Machine learning is not a technology invented in the 2010s. Its intellectual foundations are much older, and it is itself a branch of artificial intelligence.
However, advances in computing power, access to large datasets, and improvements in algorithms made machine-learning applications increasingly commercially practical during that decade.
Businesses began deploying predictive systems to identify fraud, forecast demand, recommend products, evaluate risks, and optimize operations.
Much of this technology proved useful.
But promotional language frequently expanded faster than managerial understanding.
A company could describe a statistical forecasting tool as an advanced artificial intelligence engine. A dashboard with conventional business rules might be presented as predictive intelligence.
The ability to use the terminology became easier than the ability to deliver measurable financial results.
Gartner’s 2012 Big Data Hype Cycle described big data as one of the most heavily promoted terms in the market.
What survived was not necessarily the terminology.
What survived were valuable analytical capabilities integrated into ordinary business operations.
Banks still need fraud detection. Retailers need demand forecasting. Manufacturers need predictive maintenance.
The public excitement surrounding a technology may fade while its practical usefulness continues to increase.
That is one reason the artificial intelligence investment cycle should not be mistaken for the complete history of AI.
Generative AI is a major development within a much longer technological evolution.
8. Artificial Intelligence, 2022–2026: The Most Consequential Test Yet
The public release of ChatGPT in November 2022 accelerated a new era of commercial enthusiasm around generative AI.
Businesses could now interact directly with models capable of drafting text, generating computer code, analyzing documents, summarizing information, and performing increasingly complex tasks.
Unlike some previous waves, the practical demonstration was immediate.
Users did not have to imagine every potential application.
They could experience many of them.
That is a legitimate reason for extraordinary interest.
However, the scale of capital deployment raises serious questions about the relationship between technological achievement and financial return.
According to the OECD’s February 2026 research, global venture capital investment in AI companies reached approximately $258.7 billion in 2025, representing 61% of global venture capital investment by value under its methodology.
Stanford University’s 2026 AI Index also found that U.S. private AI investment reached approximately $285.9 billion in 2025 and that organizational adoption had become widespread.
These figures measure different categories of investment, but both illustrate the remarkable scale of the AI boom.
The financial results are less conclusive.
McKinsey’s August 2026 global survey found that 80% of respondents reported personal productivity improvements from AI use.
Yet only 37% attributed any positive enterprise-level earnings-before-interest-and-taxes impact to AI. Just 6% met McKinsey’s stricter definition of AI high performers.
These are survey results, not audited measurements of the entire global economy. Nevertheless, they expose a critical distinction.
Organizations are becoming better at using AI faster than they are becoming better at earning money from it.
That gap may narrow. It may even narrow dramatically.
But for investors financing enormous future expectations, the timing matters.
In June 2025, Gartner forecast that more than 40% of agentic AI projects would be canceled by the end of 2027, citing rising costs, uncertain business value, and inadequate risk controls.
That was a forecast, not an observed failure rate.
Still, it highlights the difference between an impressive demonstration and a reliable, commercially sustainable production system.
A technology can attract millions of users without producing sufficient margins for every company offering it.
And an industry can grow rapidly while many of its participants destroy shareholder value.
The Warning Has Reached the International Monetary Fund
The possibility of an AI investment correction is no longer confined to skeptical investors or technology commentators.
In an October 13, 2026 assessment, the International Monetary Fund identified an AI-driven market correction as one of several significant risks to global financial stability.
The IMF specifically highlighted the danger that weaker-than-expected AI investment returns could challenge optimistic valuations.
It also examined rising leverage and interconnected financing arrangements within the AI industry.
These concerns do not establish that a crash is inevitable.
But they demonstrate why the sustainability of the AI investment boom has become a serious macroeconomic issue, not merely a debate about technology-stock prices.
Research: OECD (2026); Stanford AI Index (2026); McKinsey State of AI (2026); Gartner (2025); IMF Global Financial Stability analysis (October 2026).
9. The Recurring Pattern: Technology Creates Value. Exuberance Misprices It.
Economic historian Carlota Perez provides an important framework for understanding these recurring episodes.
In her work on technological revolutions and financial capital, Perez distinguishes periods of rapid technological installation from the subsequent deployment of technology throughout the productive economy.
Major innovations often attract capital before their commercial structures have matured.
Financial speculation can accelerate investment and infrastructure construction. But expectations may eventually exceed what the real economy can support.
A correction can then separate speculative investment from sustainable deployment.
Not every technology trend represents a full-scale technological revolution, and not every episode follows the same timetable.
Nevertheless, the distinction between financial enthusiasm and productive application is powerful.
I would extend that distinction into four recurring gaps that executives and investors should examine.
The Innovation-to-Income Gap
First is the capability gap: A technology can perform a task in a demonstration but fail to deliver consistently under real operating conditions.
Second is the adoption gap: People may use a product enthusiastically without being willing to pay enough to support its development and operating costs.
Third is the profitability gap: Revenue can grow while computing expenses, sales costs, support obligations, and competitive price pressure prevent sustainable profits.
Fourth is the capital-timing gap: Investors may finance future earnings at valuations that assume those earnings will arrive much sooner—or become much larger—than ultimately proves realistic.
These gaps explain how technological progress can coexist with financial disappointment.
Consider a hypothetical AI application that attracts 100,000 paying users.
That sounds impressive.
But if customer acquisition, computing charges, customer support, product development, and overhead consume more cash than those subscriptions generate, scale may simply increase the company’s losses.
A business cannot solve negative unit economics merely by generating more transactions of the same unprofitable kind.
The better question is not how quickly the customer base is growing.
It is whether the business becomes economically stronger as the customer base grows.
That was true in 2000.
It remains true in 2026.
10. Why I Expect an AI Investment Shakeout by 2030
My forecast rests on five structural pressures.
Too Many Products Competing for Similar Problems
The barriers to creating basic AI-enabled applications have fallen.
Entrepreneurs can use commercially available models and development frameworks to launch applications without building foundational AI systems themselves.
This is good for innovation.
It is also a challenge for long-term differentiation.
Many AI tools offer overlapping functions: drafting documents, summarizing meetings, generating marketing content, analyzing spreadsheets, or providing general-purpose assistance.
If customers can obtain similar capabilities from several suppliers, competition will increasingly shift toward price, reliability, integration, distribution, and trust.
Some providers will establish defensible businesses.
Others will discover that having an AI feature is not a lasting competitive advantage.
Falling Technology Costs May Destroy Some Business Models
Improvements in AI efficiency can make applications cheaper and more widely accessible.
That benefits users.
But lower delivery costs can also intensify competition and reduce the prices customers are willing to pay.
An application commanding a premium subscription today may face cheaper or bundled alternatives tomorrow.
The same technological progress that expands the market may erode the margins of companies already operating within it.
This creates a paradox.
AI can become more successful as a technology while certain AI companies become less successful as businesses.
Large Platforms Can Absorb Smaller Applications
Major software and cloud platforms already control important relationships with business customers.
They can incorporate AI capabilities into products organizations use every day.
A standalone startup offering one narrow function may struggle when that same function becomes an ordinary feature of an established platform.
This does not mean small AI companies cannot succeed.
Specialized data, industry knowledge, regulated workflows, exceptional service, and superior integration can create meaningful differentiation.
But businesses built primarily around convenient access to someone else’s AI model may be vulnerable.
Infrastructure Investment May Outrun Profitable Demand
Advanced AI systems require enormous physical and financial investment.
That includes chips, data centers, energy, networking equipment, and supporting infrastructure.
Some of this investment may produce decades of economic value.
Yet long-lived assets introduce financial risks when investment is based on optimistic projections of future demand.
If revenue growth slows, financing costs rise, or utilization disappoints, investors may reassess expected returns.
The important question is not whether computing infrastructure will be needed.
It is whether all the infrastructure financed at today’s prices will earn an adequate return.
Railway history offers a warning. So does the telecommunications investment boom of the late 1990s.
Venture Capital Will Eventually Demand Different Evidence
Venture capital is designed to finance uncertainty.
But venture capital is not designed to finance unlimited losses indefinitely.
As companies mature, investors must increasingly evaluate retention, revenue quality, gross margins, cash requirements, and realistic exit opportunities.
Capital can remain plentiful for successful companies while becoming scarce for weaker competitors.
That distinction matters more than a broad declaration that funding has ended.
I expect investors to become more selective and to direct capital toward businesses that demonstrate measurable economic value.
Some funding will continue to support AI infrastructure and applications. Some will migrate toward adjacent fields such as robotics, autonomous systems, biotechnology, and other emerging technologies.
Which theme attracts the next speculative frenzy is unknowable today.
But history provides little reason to assume that investor attention will remain concentrated on the same technology indefinitely.
11. What the 2030 Reckoning May Actually Look Like
When I say the AI bubble may burst by 2030, I do not necessarily mean a single dramatic stock-market collapse comparable to March 2000.
There are several possible forms of correction.
The first is a startup shakeout: companies with weak economics exhaust their financing and shut down, merge, or sell their intellectual property.
The second is valuation compression: otherwise viable companies continue operating but become worth much less than investors previously assumed.
The third is industry consolidation: larger companies acquire specialized capabilities while customers reduce the number of overlapping AI subscriptions.
The fourth is an infrastructure correction: certain computing and data-center investments experience lower-than-expected returns, excess capacity, or financing difficulties.
The fifth is capital rotation: investors reduce exposure to undifferentiated AI businesses and redirect funding toward other emerging opportunities.
These outcomes could occur together or independently.
My base-case scenario is a selective but significant shakeout during the remainder of this decade.
| Period | My expected market evolution |
| 2026–2027 | Continued experimentation, spending, competition, and increasing pressure to demonstrate reliable returns |
| 2027–2028 | Greater differentiation between financially sustainable applications and products dependent on repeated fundraising |
| 2028–2030 | More consolidation, failures, repricing, and reallocation of capital toward proven business models and emerging technologies |
This timeline is a judgment-based scenario, not a statistical prediction.
There is no reliable historical formula that can identify the precise date a speculative investment cycle will reverse.
It is entirely possible for technological progress to support high valuations longer than skeptics expect.
It is also possible for corrections to begin sooner.
Nevertheless, my central expectation remains clear:
By December 2030, I expect a substantial portion of today’s independent AI application market to have disappeared, consolidated, or been forced to fundamentally change its business model.
The survivors may become exceptionally valuable.
12. Why My Prediction Could Be Wrong
A credible business forecast must explain the conditions under which it could fail.
There are strong arguments against my 2030 thesis.
First, AI capabilities could continue improving fast enough to generate new sources of commercial demand that investors currently underestimate.
Second, enterprise adoption could translate into substantially higher productivity and earnings as organizations redesign their workflows rather than merely add AI tools to existing systems.
Third, declining computing costs could enable entirely new applications and business models, making the market considerably larger than today’s estimates.
Fourth, much of the AI infrastructure boom is being driven by companies with substantial existing businesses, revenues, and financial resources. That differentiates portions of today’s market from the weakest dot-com-era companies.
Fifth, the current AI ecosystem contains genuine operational capabilities with immediate commercial uses. Its adoption is not based solely on an imagined future.
These are meaningful differences.
They could produce a gradual rationalization rather than an abrupt collapse.
But they do not eliminate the risk of overvaluation.
In fact, one of the most dangerous investment assumptions is that a technology’s eventual success guarantees attractive returns at any entry price.
It does not.
A company can become extraordinarily important to the economy and still disappoint investors who paid too much for it.
That is why I distinguish my confidence in the future of AI from my confidence in the future of every AI investment.
The former is high.
The latter is substantially more selective.
13. What Business Leaders Should Do Before the Market Corrects
The appropriate response to speculative excess is not to reject technological innovation.
It is to impose stronger economic discipline on investment decisions.
I propose six questions for every board, CEO, CFO, venture capitalist, and technology entrepreneur evaluating an AI opportunity.
1. What specific economic problem does the technology solve?
A product should address an identifiable business need. Novelty, convenience, and impressive demonstrations are not substitutes for a valuable problem.
2. What measurable financial benefit does it create?
Identify whether the application generates incremental revenue, reduces avoidable costs, improves margins, reduces losses, or releases productive capacity.
Where the benefit is not immediately financial, establish credible intermediate measures rather than inventing an ROI figure.
3. Would customers continue paying if the excitement disappeared?
Paid retention, renewal behavior, and willingness to pay are stronger evidence than registrations, downloads, or occasional usage.
4. Does the business possess a durable advantage?
Evaluate proprietary data, workflow integration, industry expertise, switching costs, distribution, customer relationships, and service reliability.
A feature that competitors can reproduce quickly is not necessarily a business moat.
5. Can the business survive a less generous financing environment?
Examine cash burn, gross margins, infrastructure commitments, and the path to positive operating cash flow.
Stress-test the business against slower sales growth, falling prices, higher costs, and delayed fundraising.
6. Is management measuring outcomes or celebrating activity?
Number of models deployed, prompts processed, agents created, and employees trained may be useful operating indicators.
They are not, by themselves, proof of economic return.
The board should demand evidence that technology spending improves the company’s long-term economics.
These questions apply to AI investments today, but they are equally useful for evaluating whatever technological revolution receives the next wave of enthusiasm.
14. The Next New Economy Is Already Being Invented
Technological revolutions do not wait for financial markets to finish correcting the previous one.
New breakthroughs emerge while earlier industries are still maturing.
Robotics, autonomous physical systems, quantum technologies, synthetic biology, advanced materials, and new energy technologies could all attract substantial investor interest.
Some may create entirely new markets.
Others may disappoint.
The important point is not to predict which industry will become fashionable next.
It is to recognize that capital repeatedly moves toward opportunities offering the prospect of extraordinary returns.
This is not inherently irrational.
Entrepreneurial investment is essential to economic development. Without investors willing to finance uncertain technologies, many transformative innovations would never reach commercial scale.
The problem emerges when markets begin treating possibility as certainty.
At that point, the valuation of a company may become more dependent on the narrative surrounding the technology than on the economics of the enterprise itself.
Such periods often produce valuable innovation alongside substantial financial waste.
The challenge for capital markets is to finance enough experimentation to discover the future without treating every experiment as a future monopoly.
That balance is difficult.
History suggests we repeatedly fail to maintain it.
15. The Most Dangerous Sentence in Business History
There is one claim that appears, in different forms, during virtually every major episode of speculative enthusiasm:
The old rules no longer apply.
Railway promoters had versions of it.
Internet investors had versions of it.
Digital evangelists, blockchain enthusiasts, and proponents of every transformative technological movement have expressed similar confidence.
Today, some AI advocates suggest that the extraordinary capabilities of intelligent systems make conventional business economics less relevant.
I believe the opposite.
The greater the technological opportunity, the greater the need for financial discipline.
Extraordinary innovation does not eliminate the cost of capital.
Rapid adoption does not eliminate competition.
Large addressable markets do not guarantee attractive margins.
Revenue growth does not guarantee positive cash flow.
And a company cannot indefinitely finance an economically weak business model simply by attaching itself to the world’s most fashionable technology.
None of these observations diminishes the transformative potential of artificial intelligence.
They establish the conditions under which its economic benefits can become sustainable.
Conclusion: The Future Belongs to the Survivors, Not the Loudest Promoters
Tulip Mania did not end the flower trade.
Railway Mania did not end rail transport.
The dot-com collapse did not end the internet.
The digital transformation movement did not end the need for technology-enabled business improvement.
IoT’s inflated forecasts did not eliminate industrial connectivity.
Blockchain’s speculative excesses did not eliminate every possible application of distributed ledgers.
And the eventual AI investment correction—if my forecast proves correct—will not end artificial intelligence.
It may do something more important.
It may separate businesses that deliver measurable economic value from businesses that merely sell the promise of it.
The great opportunity is not necessarily to identify the most celebrated technology of the decade.
It is to identify the durable business models that remain valuable after the celebration ends.
That requires distinguishing technological capability from customer value, customer value from company profit, and company profit from the price investors pay for ownership.
Those distinctions may sound ordinary.
They are also the distinctions that speculative markets repeatedly overlook.
I expect artificial intelligence to become a permanent and increasingly important part of the global economy.
I also expect much of the current AI investment landscape to look very different by 2030.
Some companies will disappear. Others will be absorbed into larger platforms. Still others will survive by specializing, reducing costs, improving reliability, and solving business problems more effectively than their competitors.
A smaller group may become the defining enterprises of the next economic era.
Meanwhile, venture capital will continue looking for the next frontier.
It always does.
The fundamental mistake would be to confuse the disappearance of speculative enthusiasm with the disappearance of technological progress.
History shows that the two frequently move in opposite directions.
The greatest technological revolutions do not require every investor to win. They require the underlying innovations to survive.
And therein lies my prediction for 2030:
Artificial intelligence will survive. Many AI businesses will not. The financial hype will eventually fade, but the economic transformation will continue.
The next new economy will arrive with another extraordinary promise.
Investors will declare that the world has changed forever.
They may be right.
But the laws of economics will still be waiting for the financial statements.
Research Sources and References
The historical findings and contemporary statistics in this article are drawn from the following academic, institutional, and industry sources. The 2030 forecast and strategic framework are the author’s analysis, not conclusions attributed to these institutions.
1. Anne Goldgar (2007). Tulipmania: Money, Honor, and Knowledge in the Dutch Golden Age. University of Chicago Press. Publisher’s research overview.
2. Lorraine Boissoneault (2017). There Never Was a Real Tulip Fever. Smithsonian Magazine. Read research discussion.
3. Federal Reserve Bank of New York (2015). Crisis Chronicles: Railway Mania, the Hungry Forties, and the Commercial Crisis of 1847. Read historical analysis.
4. J. Bradford DeLong and Konstantin Magin (2006). A Short Note on the Size of the Dot-Com Bubble. National Bureau of Economic Research, Working Paper 12011. Read paper.
5. IEEE Spectrum (2016). Popular Internet of Things Forecast of 50 Billion Devices by 2020 Is Outdated. Read analysis.
6. U.S. Securities and Exchange Commission (2017). Investor Bulletin: Initial Coin Offerings. Read bulletin.
7. Gartner (2012). Hype Cycle for Big Data, 2012. Research summary.
Carlota Perez (2002). Technological Revolutions and Financial Capital: The Dynamics of Bubbles and Golden Ages. Edward Elgar. Author’s research and publications.
8. OECD (February 2026). Venture Capital Investments in Artificial Intelligence Through 2025. Full report.
9. Stanford Institute for Human-Centered Artificial Intelligence (2026). The 2026 AI Index Report. Full report and economic data.
10. McKinsey & Company (August 2026). The State of AI: Global Survey 2026. Read findings.
11. Gartner (June 2025). Gartner Predicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027. Read forecast.
12. International Monetary Fund (October 13, 2026). Global Financial Markets Face Rising Risks From AI, Bond Yields, and Cyberattacks. Read IMF assessment.
Author’s position: This article is an independent assessment of technological investment cycles, historical patterns, and emerging financial risks. Its forward-looking conclusions are analytical judgments, not guarantees or investment recommendations.
Author: John S. Morlu II, CPA is the CEO and Chief Strategist of JS Morlu LLC, a licensed public accounting and management consultancy firm in Woodbridge, Virginia. He has more than 20 years of professional experience in auditing and advisory work, including service as Auditor General of Liberia and FAR and DCAA compliance work at Unisys Federal Systems. He has led international audit engagements and works across the firm’s government contracting, assurance and advisory practices. He is also the founder of ReckSoft, FinovatePro, Fixaars, Signal Playbook AI and Ratevora.