The AI Investment Bubble vs. Real Businesses
Why hype wins the capital while practical, problem-solving companies fight for every customer
Primary keyword: AI investment bubble | Secondary: customer acquisition, small business growth, innovation resistance, dot-com bubble
There is a peculiar irony in today's economy. A three-person startup with a polished demo and "AI" on the first slide of its pitch deck can raise millions before it has a single paying customer. Meanwhile, a profitable company that solves a genuine, local problem, such as cold-chain delivery for rural clinics or water-saving irrigation for a dry farming region, can spend years earning each new client. Understanding the AI investment bubble is no longer a concern reserved for venture capitalists. It shapes who gets funded, which ideas get a hearing, and how hard it is for everyone else to be noticed.
This guide is written for the businessperson weighing an opportunity, the student learning to read markets, and anyone curious about why money and attention flow where they do. If any of the following sound familiar, you are in the right place:
Your product clearly works, yet prospects keep saying "let me think about it."
Investors ask whether you "use AI" before they ask about margins or customers.
Your marketing budget buys fewer customers every year, even as your product improves.
You feel pressure to put money into AI stocks or startups because everyone around you seems to be doing it.
Good ideas in your community stall because people prefer what they already know.
If your looking for an escape and are looking for melodic EDM music with that European/Mediterranean/Balkan sound and you need a track for summer, traveling, self development, and overcoming personal adversity then listen to Bay of Kotor. It has accordion, chords, orchestral chords, guitar and more.
When creating music, its about crafting a quality song that does not follow a trend but something that represents change through timelessness.
Listen/Follow Pavł Polø on Spotify.
GOLD NUGGET Hype raises capital; trust wins customers. The two run on different clocks, and the AI investment bubble has widened the gap between them. |
AI Investment Bubble
1. Why the Money Follows the Hype
Start with the numbers. According to PitchBook-NVCA data, AI companies captured roughly 65% of U.S. venture deal value in 2025, and about half of all deal value went into a tiny fraction of mega-rounds for firms such as OpenAI, Anthropic and xAI (Truebridge Capital, 2026). The concentration intensified in early 2026: Crunchbase data show AI took roughly 80% of the record ~$300 billion in global venture funding in Q1, with just four labs absorbing close to two-thirds of it (AI2Work, 2026). Law firm Gunderson Dettmer (2026) reports that non-AI companies faced a noticeably harder fundraising environment, and that ten companies received more than half of all U.S. venture dollars in 2025.
So how do smart people get pulled into the AI investment bubble without evaluating the real opportunities around them? Several well-documented forces are at work.
The psychology of following the crowd
Information cascades. Economists Bikhchandani, Hirshleifer and Welch (1992) showed that once enough people take the same action, newcomers rationally copy them and stop weighing their own private information. A crowd can therefore be confident and wrong at the same time.
Money chasing deals. Gompers and Lerner (2000) found that when capital floods into venture funds, the valuations of new investments rise, independent of the quality of the companies being funded. More money does not mean better businesses; it often just means higher prices.
Stories beat spreadsheets. "We are building the future of intelligence" is easier to repeat at dinner than "we run a profitable regional logistics firm with 14% margins." Narrative travels faster than evidence.
The label sells, even when it is false. In March 2024, the U.S. Securities and Exchange Commission settled its first "AI washing" cases against two investment advisers, Delphia and Global Predictions, for overstating their use of AI; they paid $400,000 in combined penalties (O'Melveny, 2024).
What gets overlooked when the AI label dazzles
Unit economics: does each customer generate more money than it cost to win them?
Proof of demand: paying, returning customers rather than pilot announcements.
Defensibility: could a larger platform copy the product next quarter?
Real problems with real budgets: logistics, healthcare access, agriculture, housing and local services rarely trend online, yet they pay reliably.
2. Echoes of 1999: Is This the Dot-Com Bubble Again?
The comparison is fair in several ways. Capital is extraordinarily concentrated. Even industry leaders acknowledge the froth: in August 2025, OpenAI CEO Sam Altman told The Verge he believed investors were overexcited about AI and said the moment reminded him of the dot-com era (TipRanks, 2025). And the payoff remains uncertain for many buyers. A widely cited 2025 report from MIT's Project NANDA found that 95% of organizations it studied saw no measurable return from generative-AI initiatives, despite an estimated $30 to $40 billion in enterprise spending (Ramel, 2025). That figure deserves caution: critics, including Wharton professor Kevin Werbach, questioned the transparency of its methodology (BankInfoSecurity, 2025). Still, the core finding, that pilots often stall before they reach production, is consistent with what many executives report.
There are real differences, too. Today's AI leaders are among the most profitable companies in history, and much of their spending is funded by operating cash flow rather than speculative stock issuance. The fairest conclusion is that the AI investment bubble looks like a bubble in parts, especially in private valuations and copycat startups, rather than a fiction in whole.
History offers a useful lesson. The Nasdaq Composite peaked on March 10, 2000, then lost about 78% of its value by October 2002 and did not regain that peak until 2015 (Financial Advisor, 2022). Yet research by Goldfarb, Kirsch and Miller (2007) found that roughly 48% of dot-com firms survived at least five years, comparable to early automobile and television makers, and that survival was unrelated to how much private equity funding a firm received.
GOLD NUGGET In the dot-com era, the internet was real and the prices were not. Survival had little to do with how much money a startup raised. Expect the AI cycle to sort winners and losers the same way. |
3. The Small Business Reality: Customer Acquisition Hurdles
While capital crowds into AI, ordinary businesses face a steeper climb. Federal data show that roughly one in five new establishments closes within its first year, and only about half survive five years (U.S. Bureau of Labor Statistics, 2024). The single most common bottleneck is not the product; it is getting enough of the right customers, fast enough, at a sustainable cost.
Rising acquisition costs. SimplicityDX (2022) estimated that online merchants lost about $29 on each new customer, up from $9 in 2013, with acquisition costs up as much as 60% over five years. Apple's iOS 14.5 privacy changes, the decline of third-party cookies and new privacy laws made targeted ads less precise and more expensive.
Paying to be seen. Free social reach has largely disappeared (see Section 7), so visibility now carries a price tag.
A more crowded field. Americans filed a record 5.67 million business applications in 2025, more than double the roughly 2.5 million filed in 2005, although many never become employers (U.S. Census Bureau, 2026).
A capital drought for non-AI firms. Investors' focus on a few AI giants leaves less patient money for regional or niche solutions (Gunderson Dettmer, 2026).
No track record. A new company has no reviews, no referrals and no recognizable clients, which is exactly what cautious buyers look for.
Geographic ceilings. A business built around a local need serves a finite market and cannot simply buy national scale.
4. Why Good Ideas Meet Closed Minds
Even when a solution is objectively better, people often resist it. Samuelson and Zeckhauser (1988) documented status quo bias: decision-makers disproportionately stick with the current option, even when alternatives offer more value. Ram and Sheth (1989) went further, identifying five barriers that drive innovation resistance:
Usage: the product disrupts established routines.
Value: the benefit does not seem worth the price or the switching effort.
Risk: buyers fear financial, physical or social downsides.
Tradition: the idea clashes with cultural norms or "how we do things here."
Image: the product, category or origin feels unfamiliar or carries negative associations.
Everett Rogers' (2003) diffusion research explains why this matters commercially: innovators and early adopters together make up only about 16% of a market. Everyone else waits for proof from people they trust. Carol Dweck's (2006) work on mindsets offers a helpful lens here. People who see ability as something that can grow treat new experiences as chances to learn, while a fixed outlook treats them as a threat. Buyers, managers and investors bring those same attitudes to unfamiliar products.
The current climate amplifies the problem. The 2026 Edelman Trust Barometer, based on nearly 34,000 respondents in 28 countries, found that 70% are hesitant or unwilling to trust people with different values, backgrounds or information sources. Richard Edelman described a public retreating toward the familiar rather than accepting the perceived risk of change (Edelman, 2026). Resistance often sounds like this:
"We've always done it this way."
"Let's wait until someone else tries it first."
"Who else around here is using this?"
"It sounds too good to be true."
5. Why a Progressive Business Environment Helps
Location changes the odds. Glaeser, Kerr and Ponzetto (2010) showed that entrepreneurship clusters strongly by region and is closely linked to a local supply of small suppliers and entrepreneurially minded people. In open, experimental markets, new ideas meet more early adopters, more tolerance for failure and faster word of mouth.
Peer effects make this concrete. Studying California solar installations, Bollinger and Gillingham (2012) found that each additional installation in a zip code raised the probability of another adoption there by 0.78 percentage points, with even stronger effects at the street level. When people see neighbors using something, the perceived risk falls. A progressive area therefore provides:
A denser pool of early adopters willing to try unproven products.
Visible local proof that accelerates the next sale.
Networks, accelerators, pilot programs and media that spread the word.
A culture where an experiment that fails is seen as learning, not disgrace.
The strategic insight: prove the idea where minds are open, then carry that evidence into more cautious markets, where competition is often thinner and loyalty, once earned, runs deep.
6. The Exposure Rule: How Many Touches Win a Customer?
Marketers often cite the "Rule of 7," the idea that a prospect must encounter a message seven times before buying. It is frequently traced to 1930s Hollywood studios, but no rigorous study establishes seven as a magic number. The research supports a more nuanced picture:
Familiarity breeds liking. Zajonc (1968) showed that mere repeated exposure increases positive feelings toward a stimulus. Bornstein's (1989) review of more than 200 experiments confirmed the effect but found it weakens with heavy repetition.
Three exposures may be enough to register. Krugman (1972) proposed that the first exposure sparks curiosity, the second brings recognition and evaluation, and the third serves as a reminder at the moment of decision.
Returns diminish. Simon and Arndt (1980) found that advertising response typically shows diminishing returns rather than a fixed threshold that must be crossed.
The practical rule: plan for several meaningful touches across different channels, and expect the number to rise with the buyer's perceived risk.
More touches needed | Fewer touches needed |
Unknown brand, no reviews | Referral from a trusted person |
High price or long commitment | Low price, easy trial |
Unfamiliar product category | Familiar category, clear benefit |
Skeptical or insular audience | Visible use by local peers |
Single repeated ad | Varied messages across channels |
GOLD NUGGET Exposure builds familiarity, but trust closes the sale. One referral from a neighbor can outperform a dozen anonymous ads. |
7. Why There Is More Resistance in 2026 Than in 2009
In 2009, the economy was in recession, but the channels to reach customers were wide open. Facebook had launched business pages in 2007, and brands could reach a large share of their followers for free. That window closed quickly: a Social@Ogilvy analysis found average organic reach fell from about 16% in 2012 to roughly 6% by early 2014, and to about 2% for large pages (Netimperative, 2014). Today, several pressures stack on top of one another:
Attention costs more. Free reach has faded, and privacy changes have made paid targeting less efficient (SimplicityDX, 2022).
Competition is denser. Record business formation means more voices competing for the same customers (U.S. Census Bureau, 2026).
Trust has turned inward. People now rely heavily on "someone like me" and are slower to trust unfamiliar sources (Edelman, 2026).
Hype fatigue. After years of bold claims, including documented AI washing, buyers discount promises and demand proof (O'Melveny, 2024).
Content saturation. Inexpensive AI tools have multiplied the volume of marketing content, so each individual message is easier to ignore.
The result is that winning a customer in 2026 takes more exposures, more proof and more local credibility than it did in 2009, even as the AI investment bubble pulls attention toward companies that may never need to win customers one at a time.
GOLD NUGGET In 2009, attention was cheap and trust was the bottleneck. In 2026, both are expensive, so the businesses that win are the ones that earn trust deliberately. |
8. Five Simple Steps Anyone Can Take
Apply the customer test before the category test. Before investing in, joining or praising any startup, ask three questions: Who pays? How much? How often? If no one can answer, you are looking at a story, not a business.
Read past the label. When you see "AI-powered," ask what the AI does that a person or a spreadsheet could not, and ask for evidence. The SEC's AI-washing cases show that the label can be misleading.
Cap your exposure to hype. Keep speculative bets to a small share of money you can afford to lose, and diversify. The dot-com crash showed that a real technology can still produce painful losses for investors who bought at peak prices. (This is general education, not personal financial advice.)
Try one new local business each month, and review it. Being an early adopter costs little, and your public review becomes someone else's trusted exposure.
Name your own resistance. When you dismiss a new idea, identify which barrier is speaking: usage, value, risk, tradition or image. Then ask whether it reflects real risk or simple unfamiliarity.
9. Business Steps for Founders and Operators
Lead with proof, not promises. Publish short case studies with specific results and testimonials from recognizable local clients.
Lower the risk of trying. Offer pilots, guarantees, starter packages or pay-after-results terms to dissolve value and risk barriers.
Win early adopters first, then cluster. Concentrate marketing geographically so visible adoption creates peer effects, as the solar research demonstrates.
Engineer touches deliberately. Plan at least three meaningful exposures across channels, such as local events, partnerships, search, email and referrals, and vary the message to avoid fatigue.
Build a referral and retention engine. The cheapest customer to acquire is the one you already have; reward referrals and invest in follow-up.
Know your numbers. Track customer acquisition cost against customer lifetime value. A commonly cited healthy benchmark is a lifetime value of at least three times acquisition cost.
Use AI as a tool, not a pitch. Adopt it where it measurably cuts costs or improves service. MIT's research found that purchased tools from specialized vendors succeeded about twice as often as internal builds (Ramel, 2025).
Partner with trust brokers. Chambers of commerce, employers, schools and community groups can lend credibility that advertising cannot buy.
Conclusion
The AI investment bubble is not a reason to dismiss artificial intelligence; the technology is real and will create durable companies. It is a reason to think independently. Capital follows stories, but customers follow trust, and trust is built through proof, proximity and patience. Whether you are investing, studying or building, the most valuable skill in 2026 may be the ability to look past the loudest narrative and recognize the quiet business that genuinely solves a problem.
Disclaimer: This article is for educational purposes only and does not constitute financial, investment or legal advice.
References
Peer-reviewed journal articles
Bikhchandani, S., Hirshleifer, D., & Welch, I. (1992). A theory of fads, fashion, custom, and cultural change as informational cascades. Journal of Political Economy, 100(5), 992–1026. https://doi.org/10.1086/261849
Bollinger, B., & Gillingham, K. (2012). Peer effects in the diffusion of solar photovoltaic panels. Marketing Science, 31(6), 900–912. https://doi.org/10.1287/mksc.1120.0727
Bornstein, R. F. (1989). Exposure and affect: Overview and meta-analysis of research, 1968–1987. Psychological Bulletin, 106(2), 265–289. https://doi.org/10.1037/0033-2909.106.2.265
Glaeser, E. L., Kerr, W. R., & Ponzetto, G. A. M. (2010). Clusters of entrepreneurship. Journal of Urban Economics, 67(1), 150–168. https://doi.org/10.1016/j.jue.2009.09.008
Goldfarb, B., Kirsch, D., & Miller, D. A. (2007). Was there too little entry during the Dot Com Era? Journal of Financial Economics, 86(1), 100–144. https://doi.org/10.1016/j.jfineco.2006.03.009
Gompers, P., & Lerner, J. (2000). Money chasing deals? The impact of fund inflows on private equity valuations. Journal of Financial Economics, 55(2), 281–325. https://doi.org/10.1016/S0304-405X(99)00052-5
Krugman, H. E. (1972). Why three exposures may be enough. Journal of Advertising Research, 12(6), 11–14.
Ram, S., & Sheth, J. N. (1989). Consumer resistance to innovations: The marketing problem and its solutions. Journal of Consumer Marketing, 6(2), 5–14. https://doi.org/10.1108/EUM0000000002542
Samuelson, W., & Zeckhauser, R. (1988). Status quo bias in decision making. Journal of Risk and Uncertainty, 1(1), 7–59. https://doi.org/10.1007/BF00055564
Simon, J. L., & Arndt, J. (1980). The shape of the advertising response function. Journal of Advertising Research, 20(4), 11–28.
Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social Psychology, 9(2, Pt. 2), 1–27. https://doi.org/10.1037/h0025848
Books
Dweck, C. S. (2006). Mindset: The new psychology of success. Random House.
Rogers, E. M. (2003). Diffusion of innovations (5th ed.). Free Press.
Data, reports and news sources
AI2Work. (2026, June 18). AI captures 80% of venture capital: Boom or bubble? (Crunchbase data). https://ai2.work/blog/ai-captures-80-of-venture-capital-boom-or-bubble
BankInfoSecurity. (2025). Why most AI pilots never take flight. https://www.bankinfosecurity.com/most-ai-pilots-never-take-flight-a-29361
Edelman. (2026, January 18). 2026 Edelman Trust Barometer reveals society slides into insularity. https://www.edelman.com/news-awards/2026-edelman-trust-barometer-society-slides-into-insularity
Financial Advisor. (2022, March 10). Tech valuations are back in focus on anniversary of dot-com peak. https://www.fa-mag.com/news/tech-valuations-are-back-in-focus-on-anniversary-of-dot-com-peak-66797.html
Gunderson Dettmer. (2026). 2025 venture capital report. https://www.gunder.com/a/web/qqWcgGFzJwPtu3ThTcyfyX/2025-venture-capital-report-final.pdf
Netimperative. (2014, March 11). Facebook ROI or RIP? Brands' organic reach is destined to die out (Social@Ogilvy study). https://netimperative.com/2014/03/11/facebook-roi-or-rip-brands-organic-reach-is-destined-to-die-out
O'Melveny & Myers. (2024, March 21). SEC targets "AI washing" in charges against two investment advisers. https://omm.com/insights/alerts-publications/sec-targets-ai-washing-in-charges-against-two-investment-advisers
Ramel, D. (2025, August 19). MIT report finds most AI business investments fail, reveals "GenAI divide." Virtualization Review. https://virtualizationreview.com/articles/2025/08/19/mit-report-finds-most-ai-business-investments-fail-reveals-genai-divide.aspx
SimplicityDX. (2022, July 19). Brands losing a record $29 for each new customer acquired. Business Wire. https://www.businesswire.com/news/home/20220719005425/en/Brands-Losing-a-Record-29-for-Each-New-Customer-Acquired
TipRanks. (2025). AI bubble fears grow as valuations mirror dot-com era. https://www.tipranks.com/news/ai-bubble-fears-grow-as-valuations-mirror-dot-com-era
Truebridge Capital Partners. (2026). State of venture capital: Investing (PitchBook-NVCA data). https://stateofvc.truebridgecapital.com/investing
U.S. Bureau of Labor Statistics. (2024). 34.7 percent of business establishments born in 2013 were still operating in 2023. https://www.bls.gov/opub/ted/2024/34-7-percent-of-business-establishments-born-in-2013-were-still-operating-in-2023.htm
U.S. Census Bureau. (2026). Business Formation Statistics (annual business applications via FRED). https://fred.stlouisfed.org/series/BABATOTALSAUS
Comments