The Artificial Intelligence Boom and the Ghost of 1929

Of course, the two speculative booms are not exactly identical, but they do share some common elements. Typically, there is great enthusiasm among investors about new technology — in today’s case, artificial intelligence — and its ability to boost the profits of companies that are able to take advantage of it. In the 1920s, commercial radio was a new and revolutionary medium: tens of millions of Americans listened to it. Sorkin points out that between 1921 and 1928, the value of shares of Radio America, Nvidia’s company at the time, rose from $1.5 to $85.5.

Another hallmark of a stock bubble is that at some point, its participants largely abandon traditional valuation measures and buy simply because prices are rising and everyone else is doing it: FOMO He rules today. By some measures, valuations were higher during the Internet stock bubble of the late 1990s than they were in the late 1920s. According to the latest report from the Bank of England’s Monetary Policy Committee, released last week, valuations in the US market are, by one measure, “comparable to the peak of the dot-com bubble.” This is true according to the cyclically adjusted price to earnings (Cape) ratio, which tracks stock prices relative to average corporate earnings over the past 10 years. If, rather than looking backwards, you focus on forecasts of future earnings, valuations become less stretched: the Bank of England report noted that they remain “below levels reached during the dot-com bubble”. But that’s just another way of saying that investors are betting on earnings growth quickly in the coming years. This is a moment when many companies have so far seen very little return on their investments in AI.

Of course, not everyone agrees that stock prices are far from reality. In a note to clients last week, analysts at Goldman Sachs said the market’s rally, largely concentrated in big technology stocks, “has so far been driven by fundamental growth rather than irrational speculation.” Jensen Huang, CEO of Nvidia, whose chips power AI systems at companies such as OpenAI, Google and Meta, said he believed the world was at “the beginning of a new industrial revolution.” However, even the authors of the Goldman report acknowledged that there are elements in the current situation that are “consistent with previous bubbles,” including large gains in stock prices and the emergence of questionable financing plans. Last month, Nvidia announced that it would invest up to a hundred billion dollars over the next decade in OpenAI, the parent company of ChatGPT, which is already a big buyer of Nvidia chips and will likely need more to support its expansion. Nvidia has said OpenAI is under no obligation to spend the money it invests on Nvidia chips, but the deal, and others like it, have drawn comparisons to the dot-com bubble, when some big tech companies were involved in the so-called “dot-com bubble.”circularTransactions that didn’t work out in the end.

Another recurring feature of the largest asset booms is outright deception, such as fraudulent accounting, the marketing of worthless securities, and plain old theft. Galbraith referred to this phenomenon as “the veil.” He noted that in difficult times creditors are stringent and audits are rigorous: as a result, “business ethics improves enormously.” In times of prosperity, creditors become more confident, lending standards deteriorate, and borrowed money is abundant. But Galbraith explained that “there are always a lot of people who need more,” and “space expands rapidly,” as it did in the late 1920s. He continued: “Just as the boom accelerated the rate of growth, the crash led to a huge increase in the rate of discoveries.

Sorkin traces the fates of Albert Wiggin and Richard Whitney, who, at the time of the accident, were, respectively, CEO of Chase National Bank and vice chairman of the New York Stock Exchange. Both men were involved in the failed efforts organized by Lamont to stabilize the market. In 1932, Wiggin became director of the Federal Reserve Bank of New York. But during Pecora’s investigation, which began that same year, it emerged that beginning in September 1929, Wiggin had secretly shorted his bank stock, using a pair of companies he owned to conduct the trades. He was forced to resign from the Federal Reserve. In 1930, Whitney, the scion of a prominent New England family, became chairman of the Stock Exchange, but was eventually exposed as an embezzler and served for more than three years at Sing Sing.

When reminded of stories like this, it’s tempting to paint the Wall Street leaders of the 1920s as a bunch of hustlers. Sorkin resists the impulse. In an afterword he wrote: “The difficulty is that, with the disgraced exception of Richard Whitney and Albert Wiggin, it is difficult to claim that any of the other major financial figures of the era did anything appreciably worse than most individuals in their positions and circumstances would have done.” Given the role played by the Wall Street elite in inflating and promoting the bubble, this is either a generous view or a worn-out commentary on fallen human nature. But in any case, it is true that speculative booms tend to take on a character of their own, creating incentives and opportunities that distort the judgment of people at all levels of the economy, from small investors and professional brokers to large corporations and financial institutions.

One aspect of the current boom that has received less attention is how it has spilled over from the stock market into the credit markets, with huge growth in so-called “private lending” by non-bank institutions, including private equity firms, hedge funds, and specialist credit companies. Last week, news organizations reported that the Justice Department had opened an investigation into the collapse of First Brands, a Cleveland-based auto parts company that, with the help of Wall Street, raised billions of dollars in murky transactions. One creditor told the bankruptcy court that up to $2.3 billion in collateral “simply disappeared,” and called for an independent examiner to be appointed. A lawyer for First Brands said the company denied any wrongdoing and attributed the collapse to “macroeconomic factors” beyond its control.

The sudden demise of a single, highly leveraged company operating in a sector far beyond the frontiers of AI may be a one-time event, with no broader implications. Or it could be a harbinger of what lies ahead. We won’t know for a while, maybe a good while. But in the words of the 19th-century English journalist Walter Paget, quoted by Galbraith, “every great crisis exposes excessive speculation in many houses which no one had previously suspected.” This time is unlikely to be different. ♦