Meanwhile, in London, Albert Grant, an indefatigable stock promoter who served as the model for the crooked financier Augustus Melmotte in Anthony Trollope’s “The Way We Live Now,” was taking advantage of the benign investment climate to issue shares in far-flung ventures of uncertain provenance, including the Lima Railways and the Lisbon Steam Tramways Company. In Germany, Bethel Henry Strousberg, whom Ahamed describes as “a mysterious Prussian railroad magnate,” built a business that employed a hundred and fifty thousand people and controlled more than fifteen hundred miles of track across Prussia, Hungary, and western Russia. In three years, some four hundred and fifty startups, many of them infrastructure-related, issued stock in Berlin, and the prices of many larger, more established companies doubled. Bank stocks were a particular favorite. Even members of the landed aristocracy, who had traditionally scorned the grubby world of finance, couldn’t resist the frenzy. “Ministers, generals, princes, and counts gamble in stocks in competition with the most cunning stock-exchange wolves,” Friedrich Engels, who was an interested observer, remarked.

It’s customary to describe this sort of speculation as a bubble, but it’s important to distinguish between unproductive bubbles, in which the objects of speculation have little or no intrinsic worth, and productive bubbles, which inflate around objects that ultimately create a great deal of lasting economic value. When the trade in exotically colored tulips collapsed in seventeenth-century Amsterdam, or when the meme-stock phenomenon went bust in 2021, nothing valuable had been created. In contrast, the railway boom and the internet stock bubble bequeathed to the economy some vital infrastructure, such as tracks, rolling stock, and fibre-optic cables.

Much has been made of the sums that A.I. labs and cloud companies such as Amazon Web Services and Microsoft Azure, which are known as hyperscalers, are spending on data centers and other A.I. infrastructure. Indeed, they are vast. According to the research firm I.D.C., worldwide outlays on servers, storage, and networking alone will be close to half a trillion dollars this year. One analysis estimates that A.I spending by the hyperscalers alone could total two per cent of U.S. G.D.P. But relative to the size of the economy the sums raised and expended on building out the railroads were even bigger. “Within three years, nearly one and a half billion dollars poured into the railroad bond market, an unprecedented commitment of capital that at its peak amounted to an extraordinary 5 percent of GNP,” Ahamed writes. “Railroad companies multiplied accordingly—by the close of 1872, more than three hundred had their securities listed on the New York Stock Exchange.”

It remains to be seen just how many A.I. startups beyond Anthropic and OpenAI go on to issue stock. But another parallel with the railroad era is already in place: the frenzy for stocks connected to A.I. has gone global. Take South Korea, home to chipmakers such as Samsung Electronics and SK Hynix. During the past twelve months, the Kospi stock index has virtually tripled, with many investors taking out margin loans to play the market or expand their positions. Last week, it emerged that this leveraged buying has proceeded so far that Korean brokerage firms have reached the regulatory limits on how much credit they can extend to their customers.

Leverage is a perennial feature of bubbles; so are warnings that they cannot last. In August, 1872, a piece in The Nation noted that out of three hundred and fifty railroad companies that filed accessible financial accounts, fewer than a hundred paid a dividend. Walter Bagehot, the editor of The Economist, warned his friends to avoid speculative investments like the ones being promoted by Grant and others. From Frankfurt, Mayer Carl von Rothschild, the head of the German arm of Europe’s largest financial house, wrote to his English partners, “The wild speculation in all the new bank shares continues to be the chief topic of conversation.” He went on to decry “all these new rubbishing schemes which absorb a good deal of money.”

Given the tendency of bubbles to last for longer than any detached observer might expect, it is the fate of Cassandras to be largely ignored—until it is too late. When the bust of 1873 did come, it began in Vienna, which was then the capital of the sprawling Austro-Hungarian Empire. During the previous few years, the cosmopolitan city had experienced a rush of speculation in real estate and stocks, much of it financed by credit. The boom created a nouveau riche who were referred to as the Ringstrasse Barons because of the palatial homes they built on the famous boulevard. On May 9, 1873—Black Friday—stocks collapsed, and fights broke out on the floor of the exchange, which was forced to close. It reopened a week later, but by then the power of leverage was working in reverse. Many speculators were insolvent, and the banks that had lent to them were in similarly dire shape. One prominent banker threw himself to his death from a high floor.

The shockwaves from Vienna quickly spread to other European financial centers, including London. For a time, the U.S. seemed immune. Over the summer, however, it became clear that the mighty Jay Cooke had overextended himself by investing large amounts of his own money in the Northern Pacific, whose transcontinental-line project had encountered costly delays. On September 18th, the New York branch of Jay Cooke & Company abruptly shuttered its doors after other banks refused to extend it any more credit. News of this development created panicked selling on the New York Stock Exchange, which was forced to close. By the time it reopened, ten days later, stock prices had gapped sharply lower, and a number of banks faced panicked depositors who desperately sought to withdraw their money.

In those days, there was no Federal Reserve standing ready to provide liquidity to stricken banks, as it did in 2008. To prevent a wholesale collapse of the financial system, the New York Clearing House, a privately run bankers’ association, organized a rescue operation in which its members agreed to pool their cash reserves and provide collateralized short-term loans to troubled banks. The tactic saved the banking industry, but the great railway boom was over, and there was no reprieve for the gullible souls who had been drawn into it. “Within five years, more than half of the $2.2 billion in railway bonds listed on the New York Stock Exchange would default, leaving investors with losses of almost $600 million,” Ahamed writes.

The upside of the bust was that about thirty-five thousand miles of track had been laid, doubling the size of the U.S. railway network. This reduced transportation costs, opened up new markets, and boosted over-all economic growth. Last year, Jeff Bezos, the founder of Amazon, argued that the A.I. boom was the latest productive bubble, or, as he termed it, “industrial bubble.” During episodes of this nature, he said, people get very excited, and every idea, good or bad, gets funded: only after the dust settles can the real winners and losers be discerned. “That is what is going to happen here, too,” Bezos said. “This is real, the benefits to society from A.I. are going to be gigantic.”

To be sure, Bezos isn’t a disinterested observer—and objections have been lodged to his argument. One is that, whereas railway lines and rolling stock can last for decades, A.I. chips and other pieces of A.I. hardware quickly wear out or are rendered obsolete, so they don’t necessarily have much lasting value. Some economists believe A.I. could also displace many workers as the benefits that Bezos mentioned accrue primarily to tech barons like him. Like Musk, Jensen Huang, Sam Altman and Dario Amodei, Bezos will not be able to disassociate himself from the consequences of the A.I. boom—or the A.I. bust, if one does come.

In the eighteen-seventies, as now, most people who bought stocks and bonds were well off. These people bore the initial brunt of the market crash. But the fallout soon affected the rest of the population. Industrial production fell and joblessness jumped. A collapse in the price of commodities, including foodstuffs, hit many farmers. “As falling prices strained the social fabric, the buoyant optimism of boom years gave way to a pervasive mood of pessimism,” Ahamed writes. Jay Cooke, who had once been lionized for his role in financing the Union victory, was attacked in the press and accused of cheating ordinary Americans out of their savings. Claiming to have been ruined, he sold a fifty-three-room mansion that he had built just north of Philadelphia and moved in with his daughter. Rumors persisted that he had salted away some of his vast wealth. Ahamed provides the kicker: “These rumors only seemed to be confirmed when, in the early 1880s, claiming to have made a quick $2 million from a Utah gold mine, he bought back his mansion and moved into his old office on South Third Street.” ♦