The dot-com bubble burst 25 years ago, AI is next?

The macroeconomic background 25 years ago is very different from today, but the outcome of each cycle bears astonishing resemblance.

The 25th anniversary is usually a day worth celebrating, but the feelings brought to investors by the bursting of the Internet bubble 25 years ago are mixed. 25 years ago, after the dazzling bull market in the Internet market peaked, the U.S. stock market entered a bear market, and the U.S. economy, which had experienced a 10-year expansion period, was dragged into recession.

The Dow Jones Industrial Average peaked in January 2000, and two months later, the S&P 500 Index and the Nasdaq Composite Index also peaked. During this period, Barron’s published a historically significant cover story titled ‘Burning Up’ on March 20, 2000, pointing out that internet companies were burning through the money poured into them by overly enthusiastic investors (or more accurately, speculators) at a frightening pace.

Immediately after its publication, the article drew a series of sharp criticisms, but Barron’s judgment proved correct. As this prescient article points out, a lot of money is invested in start-up projects that do business on the Internet, and they come to naught, along with these funds.

A familiar feeling came over me. Renowned financial historian James Grant recently wrote in Grant’s Interest Rate Observer, “Readers of a certain age will soon recall the telecom bust of the early 2000s. Today, as then from 1998 to 2002, tech companies are vulnerable to over-excitement, overordering, and overbuilding.”

The tech “Big Seven”, which generates a lot of free cash flow, is investing heavily in artificial intelligence. “Barron’s” recently reported that it is estimated that Meta Platforms, Microsoft and Alphabet will invest $200 billion in artificial intelligence this year, accounting for about a quarter of the combined revenue of the three companies. In the context of China’s launch of DeepSeek, which is able to produce AI models at a lower cost, “American exceptionalism” has been called into question, or at least the return on how much American tech giants can bring by investing heavily in AI.

Grant also pointed to the U.S. government’s dedication of all resources to artificial intelligence, especially data centers, as another warning sign. Before leaving office, the Biden administration instructed the U.S. Departments of Defense and Energy to lease land owned and managed by the federal government to expedite approvals for the construction of new data centers. The new Trump administration, along with SoftBank Group, OpenAI and Oracle, has quickly rolled out the Stargate (Stargate) project, which is expected to cost companies between $100 billion and $500 billion. This follows the U.S. government’s announcement in early 2023 of new data center construction projects that will require the combined power generation of all 94 nuclear reactors in the United States. Grant argues that this boom followed a bust that would not only shut down data center companies, but also ripple through the tech industry and credit markets as a whole.

This January also marks the 25th anniversary of the $350 billion acquisition of America Online and Time Warner, which is often seen as the peak of the Internet bubble and the most serious destruction of shareholder value in business history.

!

In January 2000, Time Warner CEO Gerald Levine (left) and AOL CEO Steve Case announced a merger that marked the pinnacle of the Internet boom.

Grant’s colleague Evan Lorenz (Evan Lorenz) believes that Musk’s offer to buy OpenAI for $97.4 billion is similar to AOL’s deal with Time Warner. He wrote in the latest issue of Grant Rate Watch that Musk may need to triple his bid because the $40 billion investment proposed by SoftBank would value OpenAI at $300 billion. Lorenz writes, “In the hottest areas of the economy, hyperscale M&A deals never appear at the bottom of the cycle.” ”

The dot-com bubble and bubble bursting are not the only precedents for the tech industry today. Radio Corporation (Radio Corp. of America) was the first “Roaring 20s” prominent tech stock, (Bryan Brian Taylor, chief economist at data provider Finaeon In a 2023 article, Taylor) wrote, “All you have to do is add the word ‘radio’ to the company name, and the stock price will soar even if there is little substance behind the company.” "The stock of the American Radio Corporation rose 200 times in the 20s of the 20th century, but in 1932 it fell 98% from its peak in 1929.

In 1986, General Electric (General Electric ) acquired the American Wireless Company at a price 72% higher than its peak in 1929, but during the same period, American consumer prices rose by over 500%. As a result, in the following decades, the stock of the American Wireless Company was a loser when calculated at its actual value, despite the far exceeding development of radio and television from the expectations of the bullish 1920s.

Indeed, the macroeconomic background 25 years ago is very different from today. Grant said that at that time, US Treasury bonds were in demand, and the US federal budget was in surplus. Now things are different, with two Middle East wars, the 2007-2009 financial crisis, and the expenditures brought by the COVID-19 pandemic, along with the expansion of projects such as Obamacare, resulting in US deficits at unprecedented levels except during wartime.

According to the forecast of the Congressional Budget Office(, the average annual fiscal deficit of the United States in the next 10 years is expected to be about 2 trillion US dollars, equivalent to 6.2% of this year’s GDP.

This means that in the coming years, artificial intelligence will compete with ‘Uncle Sam’ for funding. According to the Congressional Budget Office’s estimate, the primary driver of the deficit will be interest payments, which will account for 3.2% of GDP, higher than the 3% share of the underlying deficit (deficit after interest payments) and will not be affected by any measures from the ‘Government Efficiency Department’)DOGE(.

Grant said in an interview that while the circumstances in which we live have changed, the endings of each cycle are strikingly similar. This time, I’m afraid it’s hard to say “Happy Anniversary”.