The “Unproductive Bubble”

Unprofitable startups, small markets for new digital technologies, and little commercialization of new science

Stock and housing bubbles have been an inherent part of America’s development for centuries and many have left us with something useful. For instance, the 2000 dotcom bubble left us with rapidly growing e-commerce, web and news content, fiber optic networks, and enterprise software networks.

This is less true with the current bubble, and thus the title of this article includes the term the “Unproductive Bubble.” This article will show that the current bubble has produced few profitable startups and involved few if any new digital technologies, nor technologies involving recent scientific advances, and thus it is unlikely that much that is productive will be left once the dust settles. There is a growth in old technologies such as e-commerce but little in new technologies such as AI. The startup losses are also much larger than in the past suggesting that fewer of today’s startups will still exist in a few years than those of 20 years ago.

The “Unproductive Bubble” may not have completely popped, but the process started a few months ago. Almost every startup founded after 2005 has had its share price fall by more than 50% from its peak last year. Cathie Woods Ark Innovation Funds has fallen about 2/3 from their peak last year, the Thomson Reuters Venture Capital Index shows a 50% fall over the last few months, and Chinese indices such as the Hang Seng Tech or Nasdaq Golden Dragon have fallen even more.

Furthermore, as of May 16, 2022, there is not a single new U.S. startup in the top 100 companies for market capitalization, a situation that did not exist in previous decades, when companies such as Microsoft, Qualcomm, Google, and Facebook reached the top 100 within 14 years of their founding. More worrisome, there are only four new startups within the top 300 companies for market capitalization.

The poor performance of startups and new technologies suggests a new type of “National Development” is needed, one that enables more advances in science to be commercialized. What this new form of national development should be, is a different story — one that is being addressed by myself and other contributors to National Development — because the issue is far too complex to provide simple answers. Simply put, it is not about more government funding of R&D, which the left likes to recommend, or less government regulation, which the right likes to recommend. Those approaches have been debated and tried for many years by successive Democratic and Republican administrations (although federal R&D spending as a share of GDP has gone down) and yet things have not gotten better. Instead, this article will focus on the unproductive bubble and thus the need for big changes that go beyond the simple soundbites of liberal and conservative orthodoxy.

I begin this story with startups, particularly Unicorns, startups valued at $1 billion before becoming a public company. Second, I look at the market sizes of new technologies showing their sizes are much smaller than ones of previous decades. Third, what should we do?

Unprecedented Startup Losses

Profits are one important indicator of economic and technological growth. They signal a company is providing more value than the costs it is incurring. Value reflects the prices customers are willing to pay while costs are mostly salaries, and thus employees of unprofitable startups are not delivering as much value as the benefits they are receiving. This does not mean that profits are the only factor; but it does mean they should not be ignored. In the long run, profits will be reflected in a company’s stock price, market capitalization, salaries paid to employees, and high prices received from customers, thus signaling its contribution to society.

Whether a startup has profits at any given time is one thing, another is how much income or losses a startup accumulates over the years of its existence. This article focuses mostly on cumulative losses because they tell us more about a startup’s performance than whether they were profitable at IPO time or any other year. What is the cumulative sum of a startup’s profits or losses over time?

My analysis of the financial records of Unicorn startups, startups privately valued at $1 billion or more before they went public, show great concern. Ten now have greater than $3 billion in cumulative losses, a figure first achieved by Amazon more than 15 years ago, of which three went public in the last 18 months (WeWork, Rivian, and Robin Hood). Uber has the biggest losses ($24.5 billion) in the U.S.

Other publicly traded American Unicorns also have big losses, almost as large as the above. Of the 133 publicly traded ex-Unicorn startups that I analyzed, 23 now have greater than $1 billion in cumulative losses and another 36 have greater than $500 million. This means that 69 of the 133 ex-Unicorns, or more than half, have greater than $500 million in cumulative losses. Twenty three of the 69 went public in 2021.

A more revealing figure might be the number of ex-Unicorns with cumulative losses greater than annual revenues, a feat briefly achieved by Amazon when its cumulative losses peaked at $3 billion. There are now 89 of those ex-Unicorns among the total of 133, or about 67%, up from 60% before the Unicorn IPOs of 2021 are added.

Similar problems exist outside the United States, with many ex-Unicorns having similarly large cumulative losses in China, India, and Singapore. Video-streaming Kuaishou has the largest cumulative losses of any ex-Unicorn as of mid-2021 with $34.7 billion, about 50% higher than those for Uber. Many others have cumulative losses higher than their 2020 revenues.

Consider for a moment the plight of those startups with cumulative losses greater than 2020 revenues. Even if they were to magically achieve profits equal to 10 percent of revenue, a difficult feat, it would still take 10 years to erase the cumulative losses. Although Amazon managed to achieve this and later become one of America’s most valuable companies, few startups will likely repeat Amazon’s success, particularly when many of these 89 ex-Unicorns cited above have cumulative losses much greater than their annual revenues.

The good news is that the percent of profitable ex-Unicorns has risen slightly over the last few years. As shown in the below figure, 17 of them were profitable in the first three quarters of 2021, up from 16 in 2020 and 10 in 2019. Nevertheless, at the other extreme, the number with losses greater than 40% of revenues is not much smaller than those of previous years, suggesting that things have not gotten much better.

The biggest winners were Moderna and Coinbase. Moderna had profits of $7.3 billion on revenues of $11.3 billion in the first three quarters of 2021 largely thanks to the success of its vaccine, although many refuse to take the shots. Coinbase also did well with $2.8 billion in profits on revenues of $5.3 billion thanks to the popularity of cryptocurrencies in 2021, but not with everyone. Other ex-Unicorns with profits greater than 20% of revenues include video communication provider Zoom, e-commerce seller Etsy, and dating app Bumble. Nevertheless, the continued losses for many startups has caused stock prices to drop for biotech, real estate, green, and most of all SPAC (special purpose acquisition companies) Unicorns by at least half in 2021.

One reason for the big losses is that few of the startups commercialized breakthrough technologies, unlike startups of past decades. For instance, Silicon Valley’s name comes from the material used in semiconductors, silicon, a material used to make most transistors. Far superior to their predecessor vacuum tubes, semiconductor devices include microprocessors, memory chips, lasers, and LEDs and they are used to make computers, mobile phones, and gaming consoles. In combination with other breakthrough technologies such as glass fiber, they are used to construct the Internet.

Scientific research led to the development of semiconductors, glass fiber, and lasers and their developers won Nobel Prizes. Smaller contributions by university researchers also contributed to these revolutions, particularly ones from California’s UC system. According to Martin Kenney and David Mowery’s edited book Regional Growth, UC Berkeley developed the technologies for switched capacitor filters, analog-to-digital converters, electronic design automation, relational database software, RISC, and Arpanet. UC Los Angeles (UCLA) did so for telecom chips, UC Santa Barbara did so for compound semiconductors and instruments, UC San Diego did so for 200 startups commercializing telecom chips, and UC San Francisco did so for the ideas that led to the formation of Genentech, Chiron, and Cetus. These types of contributions are much less common today, however, and this is one reason why startups are less profitable and why we need a new system of achieving advances in science and commercializing them.

Small Markets for New Digital Technologies

Today’s new technologies have much smaller markets than the new ones did decades ago, and a lack of new technologies means much fewer opportunities for startups are being created today than in decades past. We can understand this by comparing the market sizes of past and present digital transformation technologies and their impact on past and present startups. To shorten the discussion, I mostly focus on the most successful startups, those that achieved the top 100 in terms of market capitalization.

For instance, in 1989, 21 million personal computers were sold at about $3,000 each for a total market of $63 billion or $132 billion in 2020 dollars. These PCs and the semiconductors and software within them provided opportunities for startups such as Intel, Microsoft, Compaq, Dell, EMC, and Adobe. Earlier types of computers and their software provided opportunities to Oracle and Sun Microsystems, all of whom were once within the top 100 market capitalized firms.

Bigger changes began to occur in the 1990s as networking equipment enabled these computers to be connected both within and between companies, largely based on rapid improvements in fiber optics. The Internet was born. This gave us e-commerce, websites for news and other content, enterprise software such as customer relationship management and manufacturing resource planning, and widespread use of mobile phones. These changes also quickly led to large markets. E-commerce, Internet hardware, and software, and mobile service revenues had reached $446, $315, $282, and $230 billion respectively by 2000 (1998 for mobile services), all in 2020 dollars to simplify comparisons to subsequent decades. Internet-connected personal computers also likely led to significant economic growth, with a period of high productivity gains between 1994 and 2004 that outpaced both the period from 1970 and 2004 and the period between 2004 and the present.

Many new startups grabbed these opportunities. Ones that achieved top 100 market capitalization status include Amazon, Cisco, Qualcomm, Yahoo!, eBay, Nvida, Paypal, and Saleforce, Some are still among the top 100 in market capitalization.

The 2000s were the beginning of rapid growth for smart phones, cloud computing, online advertising, social networking, and e-books. Also in 2020 dollars, cloud computing had global revenues of $127 billion by 2010, and online advertising of $81 billion by 2010. Facebook had 550 million users by the end of 2010. Some of this diffusion accelerated during the end of the 2000s and thus the quoted figures are for a year after 2010. The iPhone was introduced in 2007, the app store in 2008 and Android phones also in 2008. The global revenues for smart phones had reached $293 billion by 2012 and web browsing, navigation services, and other apps were widely used. Facebook, Netflix, and Google are three startups that benefited from these new technologies and are now among the top 100 in terms of market capitalization.

We see something shifting in the 2010s, which were a decade of growing markets for existing technologies, but less so for new ones. Although revenues for e-commerce, cloud computing, smart phones, online advertising, and other technologies mentioned above continued to grow, only one single category of new digital technology had achieved $50 billion in sales by 2020, and that was video streaming. The next closest was Big Data/Algorithms with $46 billion, tablet computers with $40 billion (iPad introduced in 2010), and OLED Displays with $32 billion in revenues. Artificial Intelligence, virtual reality, augmented reality, commercial drones, smart homes, and blockchain have even smaller markets. Even the Internet of Things had only reached 1/5 the number of connected devices projected in 2012 for 2020 and most of those devices were smart phones.

The small markets of new technologies is a big reason why Unicorn startups are mostly unprofitable. There were no big opportunities for them to grab and succeed and certainly not before the incumbents also noticed them. Those who have challenged these technologies have also done poorly. None of the virtual reality, augmented reality, smart home, drones, eVTOL (electric vertical takeoff and landing), blockchain, or AI startups have profits nor even large revenues.

For instance, consider AI and the startups who have challenged it. Despite the hype of AI and their startups, not a single Unicorn startup is profitable nor is one among the top 500 companies in terms of market capitalization If we expand the scope of AI and include Big Data and other aspects of software, the situation looks slightly better. Companies among the top 500 include Snowflake (data warehousing), Data Dog (cloud scale monitoring), and Crowdstrike (security), but none of them are even close to being profitable.

Things would likely vastly be different if self-driving cars, radiology, and other aspects of health care were more successful applications for AI. Instead, the growth in AI has been slow and much of it has been in the sectors of advertising, news, retail, and finance, sectors dominated by big companies such as Google and Amazon. Nevertheless, the small market of AI suggests that much of the claims made by the big tech companies about AI are hype, and some of them may even be lies.

This pessimistic story of startups and technologies is consistent with Robert Gordon’s 2016 book The Rise and Fall of American Growth, an academic paper entitled Are Ideas are Getting Harder to Find, Anne-Marie Knott’s 2017 book How Innovation Really Works, and an Atlantic article entitled “Science Is Getting Less Bang for Its Buck,”. Each show that innovation has slowed in some way over recent decades with the second and the third publications quantitatively showing that research productivity has fallen and the fourth showing that the importance of Nobel Prize Winning research has dropped over time with few awards for research done since 1990.

These and other publications also point to failures of science-based technologies. The rising cost of integrated circuit, crop yield, and drug development in “Are Ideas Getting Harder to Find” suggest that new semiconductor and synthetic biology (stem cells, gene editing) technologies have been slow to emerge and despite the purported benefits of falling DNA sequencing for synthetic biology. We are also still waiting for nanotechnology, superconductors, nuclear fusion, quantum computing, neurotechnology, genetically modified organisms, and really smart AI, among others. Remember Obama’s Brain Initiative of 2013, the promise of stem cells 15 years ago, Clinton’s Human Genome Project of 2000, or even Nixon’s War on Cancer starting with the National Cancer Act of 1971?

What should we do?

Misaligned incentives are partly behind the lack of startup profits and the small markets for new technologies and thus some needed changes are well known. Many VCs take one to two percent of the funding from investors as fees, meaning that they have an incentive to push their startup funds even if the startups may not have profitable exits. Then there are the profits from selling their startups onto the public markets hopefully before the public has noticed that the startups aren’t going to achieve the level of profitability promised in the prospectus. Some use the term pump and dump. Others refer to the greater fool theory; “there will always be a greater fool in the market who will be ready to pay a price based on higher valuation.” For new technologies, academic scientists are rewarded for well-cited papers, and not by their contribution to the commercialization of new technologies. As the old adage goes, you get what you pay for and the world’s science, technology, and startup system is not rewarding the right achievements.

But the problems run far deeper. First we must all better understand value creation, the sources of value creation, and the ways in new technologies emerge and provide economic benefits. To do so we need to ask better questions about the costs and benefits of new technologies. One way to do this is by comparing new technologies with old technologies. For instance, does a new technology provide more value than did online sales of books in 1995 or personal computers with word processing and spread sheet software in 1980? Unfortunately, few discussions of new technologies address these questions. Instead, the problem has grown successively worse with vague technologies such as non-fungible token, web3.0, and the Metaverse dominating the tech headlines and nary a well-defined value proposition.

More fundamentally, we need a new method of achieving advances in science and commercializing them that goes beyond the solutions being offered by the right and left. The right acts as though any technology offered by VCs and startups must be good because it comes from the free market, when the lack of profits suggests otherwise. Similarly, the left acts as though there is great value to all the papers being written when few technologies described in these papers are being commercialized.

We need new methods of developing science-based technologies, new methods that involve different actors, different rules and incentives for universities, and different goals for the entire system. We need policy makers, professors, and consultants to lead the charge in proposing new methods and a way of debating them.