Jeff Funk is a retired Associate Professor and winner of the NTT DoCoMo mobile science award for his lifetime contributions to social science research on mobile phones. His recent articles have appeared in Scientific American, IEEE Spectrum, Issues in Science & Technology, Mind Matters News, and Ozy.com. This is the second of a three-part series for Mind Matters News in which he explores the decline in profitability of companies that depend on technological advances.

In Parts 1 and 2, we looked at the fact that today’s startups are taking longer to become profitable and longer to enter the top 100 companies in terms of market capitalization than startups founded before 2000. But why? Or to paraphrase the title, why haven’t new Googles and Amazons been created?

Optimists like to claim that startups will earn profits once they stop pursuing growth. There is some truth to this argument. Many startups have pursued growth through low prices or discount coupons that can be redeemed on future purchases. But after 10 to 15 years of pursuing growth, shouldn’t we be seeing profitability? And might the reason for the price subsidies be that, without them, demand would plummet at many startups—and thus the chances of profitability would become even lower?

Another answer, albeit a simple one, is that innovation and productivity growth have slowed. Tyler Cowen made this argument in The Great Stagnation in 2012 and Robert Gordon made it in The Rise and Fall of American Growth in 2016. These issues are clearly related. Successful startups require innovations that contribute to productivity improvements. For instance, ride sharing startups such as Uber and Lyft don’t transport more people per time spent than do taxi companies. Without such productivity advantages, which create large amounts of value, how can we expect them to generate profits like those generated by Google and Amazon?

Put another way, startups must create value to capture value. Intel created value for personal computer manufacturers that used its microprocessors. Together, the ecosystem of companies, mostly startups, created value for final customers, both business and consumer ones. The huge amount of value that Intel created enable it to set high prices and thus enable its portion of the value to be quite large.

In the case of Uber, its service is marginally more convenient than that of taxi services because rides can be booked from phones. But it still uses the same types of vehicles, drivers, and roads as taxi services do, and thus overall ride sharing has created much less value than microprocessors did. And without that value, how can it set the high prices necessary for it to capture value and thus obtain profits?

Looking more deeply, today’s startups have targeted a much different set of technologies than did startups in past decades. Silicon Valley was given its name because of the large number of semiconductor companies founded there during the 1950s to 1980s, before the change to internet companies occurred in the 1990s. But the internet has matured, making many new internet-based companies comparatively low-tech. Online sales of juicers, mattresses, and exercise bikes may have been revolutionary 20 years ago. But they are sold in the same way that Amazon currently sells almost everything, and competing with Amazon is no easy task. Food delivery has not changed much either; it uses vehicles, drivers, and roads, as did previous services. Online education may deliver content differently but it is the same content and most critiques of education focus on content, not delivery. In none of these cases is the technology revolutionary.

Today’s startups have also targeted industries that have been traditionally regulated, making the challenges much larger than was the case for the opportunities startups targeted in previous decades. Taxi services are regulated out of concerns for congestion, an issue that continues to plague ride-sharing, scooters, and bicycle rentals. Fintech is trying to challenge traditional banking companies, companies that have also been heavily regulated since the Great Depression. Education startups are fighting an even more highly regulated industry, as the huge clash between public and private schools indicates.

Finally, today’s startups have targeted these regulated, low-tech industries with a strategy of rapid growth, a strategy that raises the stakes particularly when the industries, markets, and success factors are highly unknown and thus require substantial analyses and experimentation. It would be better if startups took more time to find good hi-tech opportunities, work with regulators to define possible opportunities, and experiment with various technologies, designs, and markets to find good opportunities and do this the old-fashioned way, making a profit along the way. For instance, ride sharing startups could work with regulators to increase the number of shared rides, thus complementing the shared rides of buses and trains.

The bottom line is that successful startups aren’t being founded to the extent they were founded 20 to 50 years ago. From VCs to entrepreneurs, business schools, and consultants, there is something wrong with America’s venture capital and innovation systems. Unfortunately, proposing solutions will require us first to end the hype that permeates most discussions and prevents possible reasons and potential solutions to emerge. It is only then that we can start to propose solutions that can return America’s startup system to its past glories.

Here are Parts 1 and 2: Where have all the profitable startups gone? We must distinguish between COVID-19’s devastating impact and pre-existing problems that it is making worse. The most successful startups of today aren’t as profitable as those founded 20 to 50 years ago, suggesting that something is terribly wrong with the current startup system.

Part 2: Why do today’s tech startups disappoint investors? Only 14 of the 45 recent Unicorns showed increases larger than those of the Nasdaq generally. Companies we hear a lot about, like Zoom and Beyond Meat, are not profitable compared to an earlier generation of tech startups.