Explaining an AI bubble burst and what it could mean

The AI boom

ChatGPT sparked the AI boom and broadly impacted the technology and financial markets.

Nearly every IT vendor has announced or integrated some form of generative AI assistant or workflow into its platform.

The surge in AI interest has been driven by the technology's potential to revolutionize various industries, such as healthcare, finance and manufacturing. AI promises to automate tasks, summarize information, write content and provide deep insights through data analysis.

AI can generate images and video from simple text prompts using technologies including OpenAI's Dall-E, Stability AI's Stable Diffusion and Runway. The ability to clone voices and create digital avatars -- or potentially deepfakes -- has also sparked interest, especially with privacy and security concerns. The overall potential of generative AI has led to significant investment, with venture capitalists pouring billions into AI startups and established tech giants heavily investing in AI research and development.

While OpenAI is an early pioneer in the generative AI era, it's not the only startup in this space. Beyond OpenAI, many AI startups have raised large amounts of venture capital. Other vendors include Anthropic, Cohere, Hugging Face, Runway, AI21 Labs, Glean, Jasper, Perplexity and Mistral.

AI isn't just about startups, it has led to a boom for existing, established publicly traded technology vendors as well. Among the biggest winners in the AI boom has been chip vendor Nvidia. Nvidia's graphics processing unit (GPU) technology had previously been largely seen as a tool to help gamers, and designers and for a brief time as a tool for cryptocurrency mining. With the rise of generative AI, Nvidia's AI accelerator technologies have become the foundational component of hardware to help train growing large language models (LLMs).

From November 2022 to October 2025,  Nvidia stock and overall  market value increased by a staggering amount, fueled largely by the oversized expectation for the AI's potential. In June 2024, Nvidia surpassed the $3.3 trillion mark and by October 2025, Nvidia became the first company to reach $5 trillion in market capitalization making it the most-valued company listed on a public stock exchange in the U.S.

The Magnificent Seven group of technology vendors -- which includes Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla -- have all invested in AI. All those vendors have also benefited in one way or another from a stock bump as investors have looked to chase and identify opportunities to capitalize on the AI boom.

The private AI giants showed even more dramatic growth trajectory. OpenAI was valued at $300 billion in April and nearly doubled by October 2025 hitting a frothy valuation of $500 billion.
However, there are concerns about the ROI associated with AI as companies are spending money to find future value in betting on AI to improve workflows.

Anthropic, the other major private player in foundation models, experienced similarly explosive growth. The AI safety-focused startup raised $13 billion in September 2025 at a $183 billion valuation.

Why AI doesn't make money

Despite the massive investments and high expectations, AI has struggled to generate significant profits for many organizations.

Major tech firms have spent significantly on data centers and computing hardware such as GPUs and AI accelerators essential for training LLMs and executing inference operations.

However, many current products, such as chatbots, AI-driven search features and image-generation technologies, lack clear monetization strategies. The question remains for some whether AI is merely a feature of a larger platform or a unique technology platform in its own right.

For example, Apple integrates AI into its platforms with Apple Intelligence, which is not a separate, standalone feature that generates its specific revenue stream. Microsoft Copilot is integrated into the Windows operating system and various Microsoft services, and it doesn't require users to pay separately. OpenAI, however, does sell access to AI services and selling those AI services is its primary mechanism for generating revenue.

The fear that AI won't generate revenue has intensified dramatically by 2025. An MIT study found that 95% of organizations achieved zero return on investment from $30-40 billion in generative AI spending, while S&P Global research showed that more than 40% of companies abandoned AI initiatives in 2025, a sharp increase from 17% the previous year.

There is also research that shows it is taking longer than expected for organizations to actually benefit from GenAI investments. An October 2025 Deloitte report, found that most organizations reported achieving satisfactory ROI on a typical AI use case within two to four years. That's significantly longer than the typical payback period of seven to 12 months expected for most technology investments. Only six percent reported payback in less than a year, and even among the most successful projects, just 13% saw returns within 12 months.

The fear that AI doesn't make money, however, might be unwarranted.

GenAI is in its early stages of market evolution, said Chirag Dekate, vice president and analyst at Gartner. Dekate stated that, typically, at this stage of evolution, the market tends to favor supply-side revenue generation while demand-side enterprise adoption remains nascent.

Dekate noted core supply-side innovators, including Nvidia, TSMC, AMD, and their adjacent channel counterparts, are seeing outsized revenue impact from GenAI, and the same is true for hyperscalers such as Google, Microsoft and Amazon.

"Enterprises are in the early stages of development and engineering with GenAI solutions," Dekate said. "All layers in the GenAI value chain will need to evolve for this dynamic to become broader-based, for enterprises to experience GenAI-related revenue and broader impacts."

How AI affects the stock market

The potential for businesses to increase their profits is one factor that influences the stock market. A company's potential to generate revenue in the upcoming quarter or year is just as important to its stock's valuation as its current revenue.

Investor enthusiasm for AI stocks has propelled markets to new all-time highs in 2024 and 2025, as the prospect of future gains from AI appeared to represent a new opportunity.

The Nasdaq Composite Index gained significantly since the end of 2022, and the excitement over AI technology is one of the driving factors. While leading AI technology companies exhibit strong growth potential, financial experts advise against unrealistic valuations. Wall Street's reaction to AI developments has been a mix of genuine optimism and speculative frenzy, making it difficult to predict the long-term stability of AI-related stocks.

For instance, Amazon's somewhat disappointing second quarter fiscal 2024 earnings report was partially attributed to spending on AI without significant results, leading to a corresponding drop in its stock. Similarly, Intel's shares fell after the company announced that its extensive efforts to adapt to the AI trend would cost billions and result in tens of thousands of job cuts.

Through October 2025, market reactions to AI earnings became increasingly sensitive. Meta's Q3 2025 results showed $51.24 billion in revenue, up 26% year-over-year, with strong AI-driven advertising growth. However, the stock dropped 10% as the company warned that 2026 spending would be notably larger due to AI-related expenses. Microsoft's Q1 FY2026 results beat expectations with Azure accelerating to 40% growth, but the company took a $3.1 billion hit to net income from its OpenAI equity investment.

Why this matters to executives

Whether there is an AI bubble or not has impact that extends beyond the financial markets. IT leaders are also being directly impacted. The divergence between the hype and reality of AI has multiple implications.

• Capital allocation. A report from the Yale CEO Summit in June 2025 revealed 40% of CEOs believe AI hype has led to over-investment, yet 92% of organizations plan to increase AI spending in 2025. The challenge is that organizations recognize they are spending a significant amount on AI, but fear being disrupted or displaced by a competitor that is also investing in AI.
• Investment ROI. With many organizations struggling to gain returns on their AI investments, it's critical for IT leaders to have a solid strategy. Rather than pursuing broad, exploratory AI initiatives, leadership must demand ROI-first pilots tied to specific business outcomes before capital is deployed
• Revenue model clarity. Understanding how AI will affect the organization is critical. The challenge for executives is to demonstrate value in some way. Executives must understand whether AI is a cost-reduction tool through automation and efficiency, or if it's a revenue-growth tool, and account for it accordingly.
• Risk management. With a high percentage of organizations abandoning AI initiatives in 2025, market sentiment is clearly shifting. Executives must weigh the competitive risk of under-investing in AI against the financial risk of over-investing in unproven initiatives. Big tech giants have balance sheets large enough to absorb losses on failed AI projects, but mid-market and smaller companies have significantly less margin for error.
• Stakeholder communication. As bubble concerns gain traction, boards and analysts will increasingly demand clear AI ROI metrics, forcing executives to move beyond vague promises to concrete business impact.

What causes a market bubble?

A market bubble is a metaphor describing a period of overhype and overvaluation.

A market bubble occurs when the prices of assets in the public market greatly exceed their intrinsic valuation. The bubble pricing is driven by overly optimistic projections about growth and the belief that intrinsic valuation is not an accurate measure of the opportunistic value of the asset over time.

In financial markets, bubbles are only conclusively identified in retrospect, after the bubble has already popped and prices have crashed. Historical examples include the dot-com bubble, the housing bubble and the cryptocurrency bubble. Each of these bubbles was characterized by a rapid increase in asset prices, followed by a sudden and dramatic collapse. The common thread in these bubbles is the disconnect between the market's perception of value and the underlying economic reality.

What could cause an AI bubble burst?

Several factors could contribute to an AI bubble burst, including the following:

• Unsustainable valuations. Overvalued AI companies might not have the earnings or growth potential to justify their high valuations.
• Lack of profitable revenue streams. Many AI companies have been unable to demonstrate significant revenue growth from their AI investments. Without clear monetization strategies, continued heavy spending could become unsustainable.
• Regulatory challenges. Increasing scrutiny and regulation related to AI safety, ethics and data privacy could slow down the momentum of AI development and impact company valuations.
• Economic downturn. A broader economic downturn could lead to reduced investment in AI, further exacerbating the decline in stock prices.

Opinions differ on whether there will be an AI bubble.

"It's challenging to generalize, but I don't believe we're in a bubble," said Jay Jung, founder and managing partner at Embarc Advisors. "While valuations may be somewhat elevated, they are not drastically misaligned with future growth expectations."

Gartner's Dekate noted that, in his view, using the bubble burst analogy for AI is not accurate, partly because the framing is incorrect.

Parts of the market had unrealistic expectations around a complex technology stack and how quickly returns could be experienced, Dekate said. "For these, unwarranted high expectations are now meeting the ground reality, and often perceived as bubbles," Dekate added.

Comparisons to the dot-com bubble

There have been multiple market bubble events, among the most notable in technology is the dot-com bubble.

The dot-com bubble started to inflate in 1995 with the rise of Netscape, peaking in 2000 and bursting entirely by 2002. Some financial experts are comparing the dot-com era to the modern-day AI era.

According to Jung, the key difference is that during the dot-com bubble, elevated valuations weren't backed by potential cash flow, and many companies lacked a clear business model.

"At the end of the day, valuations must be backed by future cash flow," Jung said.

In Jung's view, today, major AI companies such as Nvidia, Meta, Google and Microsoft have clear paths to monetization. However, he noted that AI startup valuations in the private market might be nearing bubble territory, which is similar to the dot-com era startups with limited barriers to entry and unclear paths to monetization.

Alan Pelz-Sharpe, founder and principal analyst at Deep Analysis, sees many similarities between the current AI bubble and the 2000 dot-com bubble. Sharpe said he believes the underlying cause of the dot-com era bust was not so much crazy software startups but rather overinvestment and unrealistic expected returns on telecom networks building out an infrastructure for the internet era to come.

"So, in that regard, even if there is a burst this time, the infrastructure will still be there for others to build on," Pelz-Sharpe said. "AI isn't going away, but hard lessons are yet to be learned as to what the best use cases are for it in terms of investment and returns."

Pelz-Sharpe said that the dot-com bubble burst demanded a restructuring or reframing of the industry, and though brutal at the time, it was needed. He added that although many dot-com firms, particularly startups, folded, midsize and large tech firms have not been discouraged and continued to innovate and grow.

"The bottom line is that rapid advances in technology don't necessarily mean rapid adoption by businesses; processes and technical debt have to be navigated, business cases need to be built, extensive business analysis needs to be undertaken, and that all takes time," Pelz-Sharpe said. "So, I expect AI to be central to enterprise software, but it may take longer to deliver a return on the initial infrastructure investment than many early investors imagined."

Editor's note: This article has been updated with AI studies and statistics for 2025.

Sean Michael Kerner is an IT consultant, technology enthusiast and tinkerer. He has pulled Token Ring, configured NetWare and been known to compile his own Linux kernel. He consults with industry and media organizations on technology issues.