""I think it's obviously a bubble, and I think it's quite a simple story," the veteran investor and GMO cofounder said during the latest episode of the "Merryn Talks Money" podcast. Grantham compared AI to the railroad and the internet, two world-changing inventions marked by early investment bubbles which, when they popped, "brought the economy to its knees for a year or two" and "everybody lost their money.""
""Contrary to common belief, "bubbles don't occur when there's some crummy idea that gets touted," Grantham said. "All the bubbles are associated with serious things, and the more serious, the bigger the bubble." Grantham emphasized the "iron law" that when an asset doubles in price, the return from holding it halves from that point forward. "If you want to have the highest market in history, you will have the lowest returns in history going forward," he said."
"Grantham, who had predicted an epic market crash four years ago, said the release of OpenAI's ChatGPT in late 2023 generated immense buzz around AI and sparked major capital outlays by companies, ultimately staving off an economic slump. The market historian said that Big Tech companies have used their dominant market positions and influence to boost their profit margins to "excessive levels," but that won't stop the bubble he sees from bursting."
AI represents a major investment bubble comparable to the railroad and internet manias, and its eventual burst could significantly damage financial markets. Bubbles typically form around genuinely transformative technologies, and the greater the promise, the larger the speculative excess. When an asset doubles in price, expected future returns from holding that asset are roughly halved. The release of ChatGPT in late 2023 triggered intense investment and temporarily forestalled an economic downturn. Dominant technology firms have used market power to lift profit margins to excessive levels. The probability that the AI boom will avoid a major bust appears very low.
Read at Business Insider
Unable to calculate read time
Collection
[
|
...
]