Artificial intelligence is quickly creeping into many parts of our jobs and lives. The technology promises to immensely improve the way we live, and companies are throwing money at it, trying to capitalize on its rapid development.The investment opportunities are also enticing. New technologies come with lofty expectations and often produce big returns for the stocks and exchange-traded funds near them. But the real long-term benefits often don’t come from the technologies themselves. Instead, it’s their ability to move the entire economy forward.That story has played out many times in the past. New technologies come along that displace less efficient ways to produce and deliver goods and services. They make life easier and improve the profitability of businesses that can effectively capture their benefits.In that regard, AI isn’t that different from many past technologies.Planes, Trains, and AutomobilesThe ability to move people and products was severely limited in the early 1800s when horses and wagons were the main form of transportation. Speeds were slow and range short as traveling long distances could be dangerous, if not deadly.Trains improved those circumstances starting in the 1820s. They moved people and products from one location to another faster, safer, and cheaper than horses and wagons. The real economic impact didn’t come from their novelty or incremental gains in tourism. Businesses that were once stuck selling to locals could ship their goods to other markets more affordably. That exposed a lot of businesses to new customers and allowed them to grow bigger and more profitable.Decades later, the internal combustion engine led to the creation of the first automobiles, which further reduced the cost and increased the speed of transportation. Like trains, they could move goods long distances faster and safer, and they didn’t require rails. Airplanes would emerge in the early 1900s as another way to efficiently move people and goods across even great distances at faster speeds.All three were game-changing technologies for the economy. Not only did they produce new things that people wanted to buy, but those things had the ability to grow businesses outside of their own industry. Each innovation unlocked massive economic growth that we still enjoy today.Get Rich?New technologies like cars, planes, and trains are exciting not only as novel products but also for their investment potential. That’s especially true early on when expectations and returns are high. Investing early and directly in new technologies can produce life-changing wealth, but it’s a high-risk/high-reward endeavor. Very few early investors get rich and stay rich. It’s always been that way.Henry Ford of Ford Motor Company and William C. Durant, founder of General Motors, were among the earliest automotive pioneers in the early 1900s. Getting in at the ground level allowed them to take advantage of easy improvements to build successful businesses. Ford developed the first assembly line, and Durant formed a conglomerate of vehicle brands and parts manufacturers.The rest is history. Ford and GM have become two of the most successful automotive manufacturers in the world. Their long-term success suggests they were great investments, but that potential was hard to spot in the early 1900s.Ford’s first two attempts at building an automotive manufacturing company ended in bankruptcy, and Durant was ousted from GM more than once for making bad decisions. Durant lost his fortune twice during the stock market crashes of 1920 and 1929 before dying penniless in 1947. If that wasn’t bad enough, hundreds of automotive competitors went bankrupt, shut down, or were acquired by others. Wealth did not come easy to either man or their investors.A similar pattern is starting to play out with stocks and ETFs tied to AI. AI models like ChatGPT, Copilot, and Gemini are just starting to make their way into the world and have already spawned numerous copycats. They’re still in the early stages of development, and expectations are high.Those expectations have translated into heady returns for AI-oriented ETFs over the past few years. Despite a lot of hype around AI stocks, only 10 “AI and Big Data” ETFs were available at the beginning of 2023. All of them easily beat the market over the ensuing three years by betting big on stocks like Nvidia NVDA, Microsoft MSFT, Meta Platforms META, and Broadcom AVGO, among others. A typical AI ETF beat the Vanguard Total Stock Market ETF VTI by almost 12 percentage points annualized since 2023.Those outsize returns didn’t come easy, and not all the risks baked into AI ETFs have been rewarded. All 10 were markedly more volatile than VTI, causing four to lag VTI after accounting for their greater risk. High volatility severely trimmed the risk-adjusted edge for others, too.AI ETFs possess other risks beyond volatility. Many of them concentrate their portfolios around some of the largest AI-related stocks. A typical AI ETF allocates around half its assets to just 10 stocks, and none of them looks like bargains. The average price/earnings multiple was 30% higher than the broader stock market at the end of December 2025.Don’t forget about the condition of the stock market. Many investors have voiced concern about the market’s overall health. Its average P/E multiple is at an all-time high, and it has never been more dependent on the largest stocks. The chart below illustrates that VTI’s stake in its 10 largest holdings has grown steadily over the past decade, representing about 35% of its portfolio at the end of 2025. Many of its biggest constituents overlap with the largest holdings in many AI-related ETFs.The broader market’s performance is already heavily dependent on the success or failure of stocks tied to AI. Holding an AI ETF just amplifies the bet on the largest and priciest stocks in the market.ProgressHistory shows that there’s a much better long-term bet to make than focusing on the hottest names. The real economic benefits of cars, trains, and planes weren’t in picking the right stocks; it was the massive economic growth that they unlocked across industries. Arguably, logistics companies like UPS and FedEx and their customers benefited more from the advent of planes and trains than the manufacturers of the planes and trains themselves.Telephones, radio, television, and the internet all followed a similar economic progression, and AI appears to be starting an ascent along a similar trajectory. It’s still an open question for what companies and industries will benefit most from this promising new technology.Once in a while, new technologies come along that offer immense potential, but they aren’t fully developed. Companies know that a lot of tinkering and investment are required to fully realize their capabilities; that’s why AI-related spending is already massive and growing. It requires trial and error, but new technologies eventually become useful, and the best become ubiquitous. Businesses adopt them because they can do more with less, grow their revenues, and expand their profits. The best technologies eventually become necessities after much of their commercial value has been unlocked. Could you do your job without an internet connection?The secondary effects are easily forgotten. Other businesses and industries can thrive by supporting the growth and development of new technologies. A lot of steel manufacturers and oil refiners benefited from the success of Ford and GM. In a similar way, energy and utility companies have enjoyed increased demand to power the warehouses of computers necessary to fine-tune AI models.Said another way, many industries prosper when a truly transformative technology comes along, not just the companies directly tied to it. The entire economy benefits. Some of the biggest winners are not necessarily the stocks most closely linked to a technology. Google didn’t invent the internet or internet search, but it found a more effective way to deliver the technology to consumers.Furthermore, not all the companies involved in AI development today will exist or be as important 10 years from now. For every Ford and GM, there were dozens of Cartercars, Deusenbergs, and Stanley Steamers that failed to survive, let alone thrive.AI is just the latest development in a long lineage of technological advancements over the past century or more. It’s still young and requires a lot of investment and tinkering, but it appears to have the same transformative potential as past technologies. It could very easily push the economy forward and improve our lives, but it remains to be seen which companies will benefit most once the dust has settled and AI is fully embedded in our lives and work. Your best bet is to diversify broadly to capture the long-term economic benefits, no matter how they show up.