Startups have been the lifeblood of America since its founding more than 200 hundred years ago. They began to grow bigger in the latter part of the 19^th century as the age of mass production approached. Companies were founded in agricultural equipment, chemicals, bicycles, electricity generation, telephone services, electrical equipment, steel, automobiles, and in many other industries, and multiple big companies emerged in each industry.

These entrepreneurial activities continued throughout the 20^th century later leading to the creation of the semiconductor, computer, and internet industries in the second half of the 20^th century. Interestingly, these companies were created after the rapid growth in America’s productivity was over, according to economics professor Robert Gordon, author of the 2016 book, The Rise and Fall of American Growth.

Nevertheless, despite the fact that these companies were founded during the beginning of America’s productivity slowdown, the years from 1975 to 2003 produced 23 that eventually spent at least two years ranked in the top 100 in terms of market capitalization.

As shown in the table below, most of them quickly became profitable, unlike today’s startups. Twenty of these companies became profitable before year 10, for which Amazon is one of the three laggards. And nine of the companies made the top 100 for market capitalization before their 15^th year.

The times they are a-changing

Things began to change in the 2000s. The percentage of startups profitable at IPO time has steadily declined since the 1980s according to Jay Ritter at the University of Florida. The most valuable ones are also taking longer to become profitable. More than 85% of the most valuable startups, which are called Unicorns because they were valued at one billion dollars or more before IPO, are still unprofitable, and this percentage doesn’t count the dozens that have been liquidated or acquired in the last few years. Most were founded before 2012 and thus have had ample time to become profitable.

Perhaps more important, but certainly related, is the fact that no company founded after Facebook in 2004 has spent more than two years in the top 100 for market capitalization.

Many readers will argue that enough time hasn’t passed for those companies founded in the last 20 years to reach their full potential. I cannot dispute that point, but there appears to be a slowdown, a subject covered in more detail in my recently published book, Unicorns, Hype and Bubbles (2024).

Number of years to reach profitability and top 100 market capitalization

One can see this change when one considers the companies that are closest to reaching the top 100. Uber has moved in and out of the top 100 over the last two years, recently pushed out by the AI frenzy. Moderna and Airbnb also had a year or so in the top 100 during the COVID-19 pandemic, the former from big vaccine sales and the latter from workers escaping cities during the years of remote work.

Currently, Airbnb, DoorDash, and Crowdstrike are the only companies founded since 2005 that are among the top 300 in addition to Uber. There are two others in the top 500, Snowflake and Block, and four additional others among the top 1000, Roblox, Zoom, Pinterest, and Carvana.

Any of these companies could possibly reach the top 100 and stay there, particularly Crowdstrike, which purportedly uses AI in its security services.

What about AI?

Many social scientists are currently ignoring the poor performance of startups because they are focused on AI and the big increases in the NASDAQ and other indices over the last two years. Most notably, the market capitalization for the Magnificent Seven — which includes Amazon, Alphabet, Meta Platforms, Nvidia, and Tesla — has risen more than $10 trillion since January 2023, albeit there have been some recent declines.

These increases followed the introduction of generative AI services in late 2022 that enable users to produce written text, images, and videos using prompts. The last two years has seen a flurry of product releases by big and small companies, the growth of AI cloud services, and the explosive growth of Nvidia, which supplies the chips that are used to power the cloud services.

Nvidia was of course already a highly successful company that has been ranked among the top 100 companies for many years through its successful graphics chips that have been inside most PCs for decades and more recently were used to process cryptocurrency transactions.

The fact that Nvidia and other members of the Magnificent Seven are the main beneficiaries of the AI boom also suggests there is something amiss about startups founded over the last 20 years. OpenAI, the startup that started the recent AI frenzy, is still far from profitability, despite being in its 11^th year of existence, one year more than Amazon, when it became profitable.

OpenAI lost about $5 billion in 2024 on revenues of $3.7 billion, and its losses are projected to continue for years. Other AI startups, from Anthropic to Mistral, Perplexity, xAI, Cohere, and Scale AI, are in worse shape than OpenAI so it is hard to be optimistic about AI changing the fortunes of today’s startups.

Many readers are probably asking why today’s startups are doing worse than those of the past, and my book discusses many reasons. In this article, I hope to convince people that we should be addressing this question. We should certainly stop making overly optimistic claims about America’s system of startups, science, and technology.