Investors Are Pouring Billions Into AI—and Ignoring the Risks

Before dawn, the parking lot outside a recently constructed data center in northern Virginia fills up. The perimeter fence is lined with contractor vans and pickup trucks, their windshields foggy from the cold. Technicians navigate lengthy server corridors inside, inspecting cables and keeping an ear out for the faint pitch shift that indicates overheating. A large portion of Wall Street’s capital is moving into these buildings, which are silently growing across the nation. It’s difficult to ignore how swiftly they appeared.

In the belief that whoever controls the infrastructure will influence the economy for decades, investors have poured enormous sums of money into artificial intelligence. Tens of billions of dollars are being spent annually by companies like Microsoft and Amazon to increase computing capacity at a rate that would have seemed reckless not long ago. One gets the impression that caution has subtly slipped out of the discourse as you watch the cranes rise and the power lines extend toward these locations.

It appears that speed is more important to investors than certainty.

Fear contributes to the urgency. No executive wants to defend their decision to wait while rivals jumped ahead. Venture capital firms like Andreessen Horowitz have invested billions of dollars in startups that promise anything from self-governing software engineers to automated legal work. From Silicon Valley conference rooms to New York trading desks, the optimism is infectious. However, optimism tends to obscure unpleasant facts.

It’s still difficult for many AI systems to produce steady profits. When experimental tools in corporate offices don’t prove to be worth the expense, they are frequently discarded. Some managers covertly acknowledge that, for the time being at least, productivity gains are still modest. However, investment is rarely slowed by those setbacks. Perhaps momentum itself has evolved into the tactic. There are lessons to be learned from history.

Older traders continue to discuss the dot-com boom, which saw low-income companies achieve extraordinary valuations. After destroying billions of dollars’ worth of wealth, the internet finally fulfilled its promise. The similarities are hard to overlook when you’re standing on a trading floor today and watching AI stocks flash across monitors. The same anxious excitement is present. The same belief that things could be different this time. And the same ambiguity.

There are even some experienced investors who have started to voice silent concerns. Given that anticipated revenues are still far below the expenditures necessary to support current valuations, analysts at Bridgewater Associates have questioned whether expectations have surpassed reality. Though they rarely make the news, those worries permeate private discussions and subtly influence choices.

Whether the math will work in the end is still up in the air.

Costs associated with the environment add yet another level of complexity. Data centers put a strain on the local infrastructure by using massive amounts of water and electricity. Nearby residents occasionally lament the continuous mechanical noise, describing a low hum that never goes away. Although they rarely show up in investor presentations, these tangible effects are a part of the true cost. After all, technology always has an impact.

The danger of technological obsolescence is another. Today’s hardware might become outdated sooner than anticipated, requiring businesses to spend even more to stay up to date. Although that cycle helps suppliers, it makes long-term returns more difficult. Those who place bets on stability might end up funding something much more brittle.

It seems like belief has turned into its own currency as we watch this happen.

Failure is not assured by any of this. Industries could still be reshaped by AI in ways that make today’s expenditures justified. Those early internet investors who survived the crash went on to make huge profits. However, a lot of other people vanished, and their wagers are only known as warning stories.

As usual, timing will determine the result.

The money continues to flow for the time being. Data centers grow. New rounds are raised by startups. Even when there is insufficient evidence, markets reward audacity. In the hopes of being early rather than late, investors keep writing checks.