In 2000, online pet retailer Pets.com spent millions of dollars on a Super Bowl commercial featuring a sock puppet. The ad became a cultural hit. Nine months later, the company collapsed. Its stock, once trading at $11, was nearly worthless.

The dot-com bubble had burst, and the puppet became a symbol of an era when investors poured money into anything ending in “.com,” regardless of whether the business could generate profit.

Twenty-five years later, Wall Street is seeing echoes of that period. This time, the frenzy centers on artificial intelligence — and the scale is far larger.

JPMorgan Asset Management reports that AI spending accounted for about two-thirds of U.S. gross domestic product growth during the first half of 2025, surpassing contributions from hundreds of millions of consumers. In other words, corporate spending on AI, not consumer demand, is doing most of the work propping up economic growth.

Such conditions are raising alarms among analysts who worry the U.S. is entering an AI-driven bubble.

What an AI bubble looks like

A financial bubble occurs when excitement inflates prices far beyond the real value of an asset. During the dot-com boom, the NASDAQ rose more than 400 percent before plunging nearly 78 percent by 2002, according to Wall Street Mojo.

The dynamic is similar today. Corporate AI investment reached $252.3 billion in 2024, Stanford data shows. Private investment grew 44.5 percent, and mergers and acquisitions increased 12.1 percent. Yet, an MIT study released in July found about 95 percent of businesses investing in AI have not made money from the technology.

Industry leaders sound the alarm

Warnings are now coming from the same figures driving the AI boom. OpenAI CEO Sam Altman has acknowledged that an AI bubble is already underway. Goldman Sachs CEO David Solomon said he expects a wave of investments that “don’t deliver returns.” Amazon founder Jeff Bezos described the environment as “kind of an industrial bubble.”

Google CEO Sundar Pichai told BBC News there is “irrationality” in the current AI surge and cautioned that no company, including Google, will be immune if the bubble bursts.

Where the billions are going

The U.S. led global private AI investment in 2024 with $109.1 billion — nearly 12 times China’s total and 24 times that of the United Kingdom, according to Stanford. PitchBook found that nearly two-thirds of U.S. deal value in the first half of 2025 went to AI and machine learning startups, up sharply from 23 percent in 2023.

Training major AI systems requires enormous capital. Google’s Gemini Ultra cost $191 million to train. OpenAI’s GPT-4 required about $78 million in hardware alone.

Some companies are also taking on massive debt. OpenAI is projected to spend $115 billion through 2029 and reportedly posted a $5 billion loss in 2024 on $3.7 billion in revenue.

A circular investment loop

Some capital flows appear to feed back into the same ecosystem. OpenAI is taking a 10 percent stake in AMD. Nvidia is investing $100 billion in OpenAI. Microsoft, a major OpenAI investor, is also a major customer of CoreWeave — where Nvidia holds equity.

Technology investor Paul Kedrosky compared this circular financing to patterns seen during the dot-com bubble.

Michael Burry, known for predicting the 2008 housing crisis, is now betting against Nvidia and argues the AI industry relies on “ridiculously small” real demand.

Why it matters

A downturn in AI could hit retirement accounts and pension funds heavily invested in tech stocks. AI-related shares have driven 75 percent of S&P 500 returns since the launch of ChatGPT. Goldman Sachs warns that a slowdown could cut the index’s valuation by as much as 20 percent.

A sharp pullback could also threaten the broader economy. The International Monetary Fund has compared the moment to the dot-com era and cautions that a correction could hinder global growth.

The dot-com parallels

Ray Dalio of Bridgewater Associates says today’s investment environment is “very similar” to the dot-com bubble. By late 2025, five companies were responsible for 30 percent of the S&P 500’s value — the highest concentration in 50 years.

When the dot-com bubble burst, the NASDAQ fell from roughly 5,000 points to below 1,000. Companies like Pets.com disappeared entirely.

Is this time different?

Unlike many dot-com startups, AI systems are widely used. OpenAI’s ChatGPT has about 700 million weekly users, making it one of the fastest-growing consumer technologies in history.

Goldman Sachs analyst Peter Oppenheimer notes that today’s AI giants are profitable, unlike many speculative companies in 2000. Major firms including Microsoft, Google, Amazon, and Meta have diversified revenue streams and large cash reserves. The “Magnificent Seven” currently trade at 27 times earnings, far below the valuations seen in 2000.

A rational technology, an irrational market

Pichai says the moment resembles the early internet: excessive investment mixed with transformative potential.

Others see unmistakable bubble conditions. “Unless you’re the most optimistic person on the planet … you know you’re in a bubble,” investor Dan Niles told CNBC.

AI may reshape the world, but that does not guarantee that every company will survive the current wave of investment. Some will endure, much like Amazon did after 2000. Others will collapse once funding dries up.

History suggests that when investors believe prices can only rise, that is usually when the fall begins — the only unknowns are when, how fast, and how painful the landing will be.