It is by no means unusual that financial analysts worry about artificial intelligence capital investments on the financial performance of firms making those investments. Such concerns continue to be raised even for successful firms including Amazon and Meta--and other firms--that are spending heavily on AI capabilities without a clear and demonstrable early financial return.
But there always are tensions between "high-risk; high-reward" gambits and "low-risk; low-reward" behaviors that can lead to "low risk; disastrous reward" outcomes. That might be particularly true in any technology-driven industry or business where technology-led disruption is possible.
In large part, that is an understandable between people who get paid to monitor quarterly financial performance and those who can--or must--take a longer view, including industry leaders, researchers, scientists or public policy advocates,
Leaders who advocate for significant investments often view it as a "once-in-a-generation opportunity" to gain a competitive edge or maintain market dominance. On the other hand, investors and financial analysts typically are wary of overinvestment without clear short-term returns.
Even within each firm, one normally expects more caution from financial executives; more enthusiasm (or at least support for higher investment) from line of business or strategy executives.
So there are recurring tensions around overinvestment, investment “bubbles” or failed implementations of new technology. In fact, in many cases, there has been at least temporary “overinvestment” in new technologies, which might be hard to distinguish from the typical “many flowers blooming” early startups to eventual consolidation we see in any new market.
Technology | Era | Period of Overinvestment | Outcomes |
Railways | 1840s | Railway Mania (1845-1847) | Market crash, numerous bankruptcies |
Electricity | 1880s-1890s | War of Currents | Consolidation of electric companies |
Automobiles | 1910s-1920s | Auto industry boom | Market saturation, Great Depression |
Radio | 1920s | Radio boom | Consolidation, formation of major networks |
Personal Computers | 1980s | PC boom | Market saturation, industry shakeout |
Internet | 1990s | Dot-com bubble (1995-2000) | NASDAQ crash, numerous dot-com failures |
Renewable Energy | 2000s-2010s | Clean tech boom | Bankruptcies (Solyndra), market consolidation |
Cryptocurrencies | 2010s-2020s | Crypto boom | Market volatility, regulatory scrutiny |
AI | 2020s-present | Ongoing AI boom | Likely to see similar patterns |
Enterprise leaders always must be alert for new technology potential to revolutionize industries, create new products or services, and allow some adopters to possibly gain a new competitive edge. They fear falling behind if they don't invest early.
Tech enthusiasts and early adopters frequently are among those who are likewise optimistic about investments in new technology, as are venture capitalists, who see the potential for high returns on investment as the technology gains traction.
But overinvestment, wasted investment and excesses also are part of the history of new technology innovation.
Venture capitalists also routinely expect they might lose money, or possibly break even or make a slight profit, on seven out of 10 such investments.
Technology | Era | Proponents' View | Skeptics' View |
Internet | 1990s | Revolutionary communication and commerce platform | Overhyped "dot-com bubble" |
Cloud Computing | 2000s-2010s | Transformative for business operations and scalability | Security concerns and loss of control |
Mobile | 2000s-2010s | New paradigm for consumer engagement | Limited functionality compared to desktops |
Social Media | 2000s-2010s | Unprecedented user engagement and data collection | Unproven monetization strategies |
Blockchain | 2010s-2020s | Disruptive for finance and data management | Speculative technology with limited practical applications |
5G | 2010s-2020s | Enabling technology for IoT and smart cities | High infrastructure costs with uncertain ROI |
Nor are those examples limited to today’s information and communication technologies. Indeed, overinvestment has been characteristic of many infrastructure-dependent technologies in the past.
technology | Era | Period of Overinvestment | Outcome |
Canals | Early 1800s | Canal Mania (1790s-1830s) | Many canals became unprofitable due to competition from railroads; financial losses and bankruptcies |
Telegraph | Mid-1800s | Rapid expansion (1840s-1860s) | Market consolidation, with many companies failing or merging; eventual displacement by the telephone |
Railroads | Mid-1800s | Railway Mania (1840s-1850s) | Market crash, numerous bankruptcies; led to more regulated and sustainable growth |
Telephone | Late 1800s | Telephone boom (1880s-1890s) | Overbuilding and financial strain; eventual consolidation into major companies like AT&T |
Internet Apps | Late 1990s | Dot-com bubble (1995-2000) | “Internet bubble” and dot-com crash |
The point is that financial analysts have reason to be skeptical of excessive investments while investors and business leaders must make bets in hopes of profiting from the innovations. Perhaps only VCs routinely expect to fail often, as part of the process. But technologists also know that up to 70 percent of technology projects fail.
And virtually all investors and startups believe they will succeed, despite the obvious historical evidence that many to most new firms will eventually fail, or be consolidated into larger entities. So even if some see AI investment as a “bubble waiting to burst,” others see a necessity to invest heavily.
Will there be overinvestment and failed implementations? Almost certainly. We might reasonably expect failure rates up to 70 percent for AI projects, as that would be expected for information technology projects overall.
But is there danger in no investment or too-slow investment? Also almost certainly, in at least some cases. Even if 70 percent of implementations do not create value, we are left with the near-certainty that up to 30 percent will do so, and that at least a few of those implementations could deliver extensive or disruptive value.
We might also note that within each large enterprise, there likewise are roles that tend to incorporate more risk, and those that seek to minimize risk.
Innovation specialists; business development roles; product managers and venture investment officials might routinely have to advocate for high-risk and high-reward investments.
Other roles lean towards risk aversion, including compliance officers; financial roles; line of business managers and legal staff.
We might tend to agree that, in any technology-driven business, the biggest risk is not taking any risk. When technology enables rapid change, perhaps the only strategy that is guaranteed to fail is “not taking risks."
Financial analysts are virtually required to raise issues about large capex or high debt to support new initiatives. Business leaders in technology-driven industries also are virtually required to make big and risky decisions rather routinely.
Financial analysts are required to assess quarterly financial performance. But business leaders of technology-driven businesses also must make big decisions of a longer-term nature that are, of necessity, risky.
It’s a tough balancing act.
