FOMO is driving investments in AI, mirroring the dotcom boom

The dotcom boom began to take shape in the late 1990s, fueled by excitement over the internet’s potential to transform business and communication. Startups with little revenue and no clear path to profitability attracted billions in investment. The prevailing logic was not about current performance but about securing a stake in what seemed like an innovative future. Investors bought shares in internet companies based on projections of dominance.

Cisco Systems, which manufactured the networking equipment powering the internet’s expansion, became a symbol of this optimism. As companies rushed to build out digital infrastructure, demand for the company’s products soared. By March 2000, its market value had peaked at $550 billion, briefly making it the world’s most valuable company, with the belief that the internet boom would continue indefinitely.

That belief collapsed within months. The bubble burst later in 2000, erasing trillions in market value. Cisco’s stock price fell more than 80%, and hundreds of internet companies shut down. Pets.com, famous for its Super Bowl ads, folded within nine months of going public. Webvan burned through $800 million before closing its online grocery delivery service. Investors lost fortunes, workers lost jobs, and optimism gave way to caution.

Yet the infrastructure built during the boom, including fibre optic cables, data centres, and an early digital ecosystem, remained. These foundations later enabled survivors like Amazon and newcomers like Google to thrive, proving that while the dotcom bubble burst, the internet’s long-term promise was real.

A similar pattern has emerged in 2025, but the pace has accelerated. Investment in artificial intelligence (AI) now exceeds the dotcom boom in both speed and scale. John Chambers, who ran Cisco during the dotcom crash, describes the current AI boom as following similar dynamics but moving five times faster. The technology promises to transform how work gets done across industries, from software development to customer service to marketing and advertising and legal analysis. Companies are spending hundreds of billions of dollars on infrastructure to train and deploy AI systems. The question mirrors the dotcom era: how much of this spending will create lasting value, and how much will be written off when reality fails to match projections?

Sam Altman, OpenAI’s CEO, told reporters in August that the AI market has entered bubble territory. The assessment came from someone whose company stands to benefit from continued investment. Altman drew parallels to the dotcom boom, noting that when bubbles form, smart people become overexcited about a kernel of truth.

Similarly, Jeff Bezos, speaking at Italian Tech Week in Turin, admitted that it’s true while claiming that it “can even be good”. Bezos pointed to the biotech bubble of the 1990s as precedent. That cycle produced important medical advances, but The Wall Street Journal reported in 2004 that cumulative losses for public biotech companies exceeded $40 billion.

According to Bezos, while this is going to happen in the AI situation as well, “the benefits to society from AI are going to be gigantic.”

On the other hand, Bret Taylor, who chairs OpenAI's board while running Sierra, an AI startup, stated that both outcomes can occur simultaneously: AI will create enormous economic value while many investors lose substantial sums. While most of the tech executives have shared their take on the bubble forming, they are positioning their companies as survivors rather than casualties.

There have been announcements of record spending. Nvidia agreed to invest $100 billion in OpenAI for data centre infrastructure. OpenAI secured a $300 billion computing contract with Oracle that exceeds the company's current annual revenue. Meta is seeking $29 billion in private capital for AI data centres. Oracle issued $18 billion in debt for AI infrastructure expansion.

In fact, Mark Zuckerberg told investors that Meta would rather misspend hundreds of billions of dollars than arrive late to artificial general intelligence. The statement reveals how fear of missing the technology shift overrides concern about near-term losses.

Stanford University's 2025 AI Index Report found that U.S. private AI investment reached $109.1 billion in 2024, nearly 12 times China's $9.3 billion. Globally, generative AI attracted $33.9 billion in private investment, up 18.7% from 2023. The money continues flowing despite mounting evidence that returns lag far behind spending.

Reports have described the mechanics of keeping the bubble inflated. Profitable AI firms invest in unprofitable ones, creating a closed loop that minimises net cash requirements. The system works until profitable companies run out of capital to deploy or investors demand actual returns. 

As executives acknowledge the risks and simultaneously bet their companies will emerge as winners when the bubble bursts, it creates a situation where rational actors collectively make irrational decisions because the cost of being wrong appears less severe than the cost of being left behind.

Investment reaches record levels despite minimal returns

Companies are projected to spend about $400 billion this year on AI infrastructure, according to reports. That figure is expected to surpass $500 billion in both 2026 and 2027. The spending has not yet produced proportional revenue. 

Reportedly, American consumers spend only $12 billion annually on AI services. A research project at MIT called Project NANDA studied AI implementation across organisations and found that 95% of organisations implementing generative AI are seeing zero return on investment despite spending between $30 billion and $40 billion. The research found that about 5% of AI pilot programs achieve rapid revenue acceleration, while the vast majority stall without measurable impact on financial performance.

Investment patterns show that funding flows to companies regardless of product readiness. Mira Murati left her position as Chief Technology Officer at OpenAI to start Thinking Machines. The company raised $2 billion at a $10 billion valuation without releasing a product or telling investors what it plans to build. An investor's pitch meeting indicated that Murati’s company would hire top AI talent but declined to answer questions about product direction or strategy. This is interesting given that the deal represents the largest seed funding round in history.

Between 2013 and 2024, the United States led global AI investment with $470.9 billion in total private funding, according to Stanford's AI Index Report. China attracted $119.3 billion, the United Kingdom $28.2 billion, and Canada $15.3 billion. Government spending adds substantial amounts to these totals. China launched a fund targeting $138 billion over 20 years. Saudi Arabia's Project Transcendence represents a $100 billion commitment. Canada allocated $2 billion for AI infrastructure.

India has become a growing market for AI investment. Data from Venture Intelligence shows AI startups raised $665 million across 109 deals in the first seven months of 2025, a 46% increase over the same period in 2024. Average deal size grew 24%, from $4.9 million to $6.1 million. The country now has more than 240 generative AI startups.

These startups face competition from international players entering with far greater resources. OpenAI's expansion into India provides infrastructure access but creates challenges for local companies. When a company that raised $40 billion at a $300 billion valuation enters a market, local startups raising single-digit millions cannot compete on pricing or customer acquisition spending. The pattern repeats across emerging markets where local entrepreneurs build businesses only to face competition from companies that can operate at losses indefinitely.

Fear of missing out drives spending 

The psychology behind continued investment centres on competitive positioning rather than immediate returns. Companies operate under the assumption that the cost of missing an industry transformation exceeds the cost of overspending during a bubble.

John Chambers told The Associated Press he expects half the Fortune 500 companies to disappear, along with half their executives. The prediction stems from his view that most corporate leaders lack the skills to adapt to AI-driven change because they learned to operate on five-year planning cycles when markets now move in 12-month cycles. This creates urgency for companies that might otherwise wait for clearer evidence of returns.

Chambers even noted that if AI moves at five times the pace of the internet revolution, jobs will be eliminated faster than new roles can be created. This creates a period of adjustment where large numbers of workers need retraining for positions that may not yet exist. Reports and analyses have indicated that AI may not be able to perform many jobs, but can convince executives to fire workers and replace them with AI systems that cannot actually do the work. When the bubble bursts and money-losing AI companies shut down their services, the displaced workers will already be gone. This creates permanent damage even though the technology failed to deliver promised capabilities.

The AI boom of 2025 mirrors the dotcom era in both scale and psychology, with massive investments fueled by FOMO, minimal immediate returns, and a belief that missing the wave is riskier than overspending. History suggests a correction is inevitable, with many companies likely to vanish while a few emerge stronger. As with the dotcom bust, the lasting question is whether today’s spending will build the foundations for change, or be remembered as another costly lesson in collective overexcitement.