The AI Bubble and Dentistry, What the Crash of 1873 Can Teach Dental Practice Owners
Artificial intelligence is already moving into the dental operatory, front desk, call center, radiology workflow, insurance process, treatment presentation, and practice management system. The question for dentists is no longer whether AI is real. It clearly is. The harder question is whether the extraordinary amount of money now being invested in AI will earn an adequate return, and whether dental practices will make the same mistake investors have made during earlier technology booms, confusing a revolutionary technology with a good investment at any price.
That is what makes Liaquat Ahamed’s new book, 1873: The Rothschilds, the First Great Depression, and the Making of the Modern World, unusually relevant. Ahamed tells the story of the railroad boom of the 1850s and 1860s, when globalization accelerated, capital crossed borders at unprecedented speed, and investors poured money into railroads, real estate, infrastructure, and government debt. Railroads were not a gimmick. They changed commerce, transportation, settlement, and productivity. Yet the certainty that railroads would transform the world encouraged investors to finance too many projects, at prices and leverage levels that could not be justified by future cash flow.
The boom broke in 1873. The Vienna stock market crashed, financial stress spread across Europe and the United States, railroad companies defaulted, banks failed, and governments that had borrowed heavily found themselves shut out of capital markets. According to the National Bureau of Economic Research, the U.S. contraction that began in October 1873 lasted until March 1879, a remarkable 65 months. That does not mean the 1870s were worse than the Great Depression of the 1930s, but it does show how a financial boom built around a genuine technology can produce consequences far beyond the original investment thesis.
The most important lesson is simple. The railroads were real, useful, and ultimately indispensable. The financial excess surrounding them was also real. Those two facts were never in conflict.
Artificial intelligence may be following a similar path. Reuters estimates that Alphabet, Amazon, Microsoft, and Meta will spend more than $700 billion in 2026, much of it on AI infrastructure. The Bank for International Settlements estimates that the five largest hyperscalers are committing more than $1 trillion to AI related capital spending across 2025 and 2026. The BIS has also modeled the current AI race as a winner take most competition in which each company may rationally believe it cannot afford to fall behind, even if the industry collectively builds more capacity than is economically efficient.
That is an important distinction for dentists because the same psychology exists at the practice level. A dentist hears that AI is transforming radiology, scheduling, claims, treatment planning, marketing, call conversion, and patient communication. Competitors are adopting it. Vendors warn that practices that wait will be left behind. The fear of missing out can quickly become a purchasing strategy. Yet the right question is not whether AI is impressive. The right question is whether a specific tool improves a specific workflow enough to justify its cost.
The strongest case against calling the current AI boom a simple bubble is that demand is real. Google Cloud recently reported 63 percent year over year growth, Microsoft Azure grew about 40 percent, and Amazon Web Services grew 28 percent. Amazon disclosed an annualized AI services revenue run rate above $15 billion and said a significant portion of future capacity already has customer commitments behind it. These are not companies with nothing more than a pitch deck and a website. They have customers, revenue, distribution, infrastructure, and enormous cash flows.
That makes the current situation very different from the weakest parts of the dot com era. It also means dentists should be skeptical of anyone who says AI is just hype. It is already generating real economic value, and many dental applications will become ordinary infrastructure in the same way digital radiography, cloud software, online scheduling, and electronic claims eventually did.
But real demand does not guarantee good economics. The most vulnerable part of the AI boom appears farther down the financing chain. CoreWeave, a major provider of AI computing infrastructure, has become a useful case study. Reuters reports roughly $50 billion in debt, very large capital spending requirements, and bond yields that recently approached distressed levels. CoreWeave also has a huge contracted revenue backlog and strong relationships with major customers, which is why it would be wrong to treat the company as proof that AI is collapsing. Its lesson is narrower. Fast growth financed with large amounts of debt can become fragile when borrowing costs rise, utilization falls, or expected revenue arrives more slowly than planned.
The International Monetary Fund and the BIS are both warning about this broader mechanism. Their concern is not that AI fails technologically. Their concern is that leverage, concentrated exposures, private credit, and circular financing can magnify disappointment. A supplier can also be an investor. A customer can also be a financing partner. An infrastructure company can borrow money to buy chips from a company that has invested in it. None of that is inherently improper, but it makes the system more interconnected. If expected returns weaken, losses can travel through lenders, suppliers, funds, and customers much faster than the original business problem might suggest.
For a dental practice owner, the practical lesson is not to predict the stock market. It is to avoid making capital decisions based on the assumption that every new AI product will become indispensable. Dentists should demand evidence at the workflow level. Does the radiology system improve diagnostic consistency or patient understanding. Does the scheduling tool reduce no shows. Does the call system convert more new patients. Does the insurance tool reduce staff time or days in accounts receivable. Does the treatment presentation software increase case acceptance without confusing or pressuring patients. Does the product reduce labor, improve quality, increase throughput, or produce revenue that can actually be measured.
If the answer is yes, the technology may deserve a place in the practice even if the broader AI investment cycle eventually crashes. A dentist did not need to own railroad bonds in 1873 to benefit from rail transportation. In the same way, a dental practice does not need to speculate on AI companies to benefit from cheaper, faster, and better AI tools.
There is also a patient communication lesson. AI will increasingly influence what patients believe they know before they enter the operatory. They will arrive with chatbot answers, automated interpretations, online treatment comparisons, and price expectations generated outside the practice. The dentist’s role becomes more important, not less. The winning practice will use AI to make information clearer while preserving human judgment, trust, empathy, and accountability. Technology can help explain a lesion, show bone loss, summarize options, or follow up after a consultation. It cannot replace the relationship that makes a patient comfortable saying yes to treatment.
The same discipline should apply to vendors. Dentists should be wary of products sold primarily through fear. Claims that every practice must adopt a platform immediately, that human staff will soon be obsolete, or that one algorithm will transform case acceptance deserve the same skepticism applied to any other capital purchase. Ask for real customer retention, measurable outcomes, implementation time, integration costs, cancellation terms, training requirements, data ownership, and what happens if the vendor disappears. A strong demo is not a return on investment.
Historical analogies also need restraint. It is tempting to map every part of 1873 directly onto 2026, railroads become AI, railroad bonds become private credit, and CoreWeave becomes the new Jay Cooke. History rarely repeats that neatly. The nineteenth century had no Federal Reserve, no federal deposit insurance, a much more fragile banking system, and a monetary system increasingly tied to gold. Modern policymakers have tools that did not exist in 1873. An AI investment bust could be severe without producing anything close to a five year economic contraction.
The balanced conclusion is therefore more useful than either extreme. AI is almost certainly a durable technological revolution. Parts of the investment ecosystem also show classic signs of excess, including enormous capital spending, intense fear of falling behind, expanding leverage, concentrated financing, and valuations dependent on extraordinary future growth. The technology can succeed while many investments fail. That is not a contradiction. It is one of the oldest patterns in economic history.
Dentists should take the same approach to AI that they should take to any expensive clinical or business technology. Do not buy because everyone else is buying. Do not reject it because markets may be overexcited. Identify the problem, measure the baseline, test the tool, calculate the return, watch the workflow, and keep the patient at the center. The best dental practices will not be the ones that use the most AI. They will be the ones that know exactly where AI creates value and where it does not. Are we using AI to build better dental practices, or simply buying into the biggest technology race of our lifetime?
Join the Conversation!