FT: History explains revolutionary cycle: Brace for a crash before the golden age of AI

Key topics:Tech giants spend $750bn+ on AI data centers, $3tn forecast by 2029Perez warns AI is in a bubble phase before any golden age emergesAI could reshape industries but faces risks of crashes and instability.Sign up for your early morning brew of the BizNews Insider to keep you up to speed with the content that matters. The newsletter will land in your inbox at 5:30am weekdays. Register here.Support South Africa’s bastion of independent journalism, offering balanced insights on investments, business, and the political economy, by joining BizNews Premium. Register here.If you prefer WhatsApp for updates, sign up for the BizNews channel here..By John Thornhill.The massive investment in artificial intelligence infrastructure currently being made probably counts as the biggest and fastest rollout of a general purpose technology in history.This year and next, Google, Amazon, Microsoft and Meta alone will spend a staggering $750bn on data centres to power their AI models, with Morgan Stanley forecasting total global spending in this area will reach $3tn by 2029. But jittery investors are increasingly asking: what returns will this immense capital outlay generate? History suggests they’re right to be nervous.There are few better scholars to put AI in a historical perspective than Carlota Perez, author of Technological Revolutions and Financial Capital: The Dynamics of Bubbles and Golden Ages. In her book, Perez identifies five great technological revolutions: the industrial revolution of the late 18th century; the steam, coal and railway revolution of the 1830s; the steel and heavy engineering revolution of the 1870s; the mass production age in the early 20th century; and the information technology revolution beginning in the 1970s. Perez sees AI as an extension of that fifth technological revolution. She also argues these revolutions follow a fairly predictable cycle. An initial installation phase results in lots of creative destruction and social disruption as industries and regions are upended. That is normally accompanied by over-investment, financial mania and stock market bubbles. Nevertheless, those bubbles are often productive, funding the building of vital infrastructure that enables the subsequent mass rollout of technology — as railways or electricity grids are built — and their broader economic benefits are realised. As for AI, we remain in that manic installation phase.That point was reinforced by a report from the Massachusetts Institute of Technology, which unsettled stock market investors this week. Researchers found that 95 per cent of the companies they surveyed were getting zero return from their investments in generative AI. Sam Altman, OpenAI’s chief executive, was hardly reassuring when asked whether there was an AI bubble. “I do think some investors are likely to lose a lot of money,” he replied.A bone-juddering crash — or several crashes — therefore seems likely before we reach any golden age of AI. “I have not seen a golden age happening without a crash,” Perez breezily tells me..Read more:.The Economist: How AI could create the first one-person ($1bn) unicorn.Even more cheerily, she adds that the bursting of the AI bubble could lead to yet bigger upheavals as capital markets are misfiring. They are now focusing more on speculative games, such as crypto, than productive investments, and global debt amounts to more than three times GDP. “This could also be a trigger for gigantic instability,” she says.But it is worth investors considering how this technological revolution may differ from previous cycles. It is certainly the first revolution to be driven by software as much as hardware. That changes some financial dynamics as massive network effects come into play. Software companies can scale up quicker and go global overnight. OpenAI’s ChatGPT is being used by 700mn people every week, less than three years after launch. But if digital globalisation increases opportunities, it also magnifies risks. Look at how China’s cheaper DeepSeek AI model rattled investors in US tech stocks.Perhaps the most intriguing difference, though, is how far today’s AI companies will themselves benefit from the financial gains they help unleash. The technology is accelerating advances in many areas: biotech, robotics and material science, for instance. AI companies could well exploit their technological advantage to become significant healthcare, drug discovery or autonomous car companies. To what extent can they morph into general purpose companies and capture the fruits of the golden age?One more important lesson, though, should be drawn from previous revolutions, Perez adds. To enter a golden age, civil society needs to shape the revolution to its own ends. So, for example, earlier politicians established antitrust agencies to tame over-mighty companies and created welfare states to soften labour market disruption.Perez argues that today’s dysfunctional financial markets, the concentration of corporate power, the surge of populism and the threat of climate change have brought the world to a new turning point. How we should respond is the subject of her next book.However, as the historian AJP Taylor once wrote about the 1848 revolutions in Europe, countries can sometimes reach turning points — and fail to turn..© 2025 The Financial Times Ltd.

FT: History explains revolutionary cycle: Brace for a crash before the golden age of AI

Related Stories

Japan’s slow AI start could be its strategic advantage: Catherine Thorbecke

Why the US's current IPO mania could signal top of stock market

AI super-apps are remaking China’s internet

US Big Tech’s fat profits conceal unsettling cashflows