{"text":[[{"start":6.3,"text":"The writer is a managing director at Panmure Liberum"}],[{"start":10.15,"text":"In December 1996, Federal Reserve chair Alan Greenspan characterised the boom in technology, media and telecom stocks as showing signs of “irrational exuberance”. Almost 30 years later, we can say the same about the AI boom."}],[{"start":25.9,"text":"But while there are similarities between the current tech boom and the one a generation ago, there is an important difference. One aspect of today’s boom is already much larger than the TMT bubble ever was. In 2025, US businesses invested almost $1.5tn in IT equipment and software. At the peak of the TMT bubble, it was $466bn or $829bn when adjusted for inflation."}],[{"start":53.7,"text":"Indeed, the US economy is growing solely because of the tech boom. I calculate that over the past four quarters, 93 per cent of US GDP growth was explained by tech investments. Even at the peak of the TMT bubble, it barely reached 60 per cent."}],[{"start":70.1,"text":"The developers of large language models such as OpenAI and Anthropic are preparing for blockbuster initial public offerings later this year to benefit from investor optimism about their growth. Meanwhile, the hyperscalers Microsoft, Alphabet, Amazon, Meta and Oracle plan to invest hundreds of billions in the next five years in data centres to provide the computing power to run these models."}],[{"start":93.85,"text":"And this is where the maths of the AI boom becomes challenging. For each of these hyperscalers, I collected the consensus estimates of analysts for the capital expenditures and revenues between 2025 and 2030."}],[{"start":null,"text":"
"}],[{"start":106.69999999999999,"text":"In these five years, capital investments are expected to rise by 20 per cent a year, a growth rate never seen before in this industry. Meanwhile, revenues are expected to grow 15 per cent annually. If we make the heroic assumption that there are no costs, then the additional revenue is the profit these companies are expected to make from their additional investments in AI data centres. Yet, even under these extremely optimistic assumptions, I calculate the implied return on investment is highly negative for all of them except Amazon."}],[{"start":139.1,"text":"These numbers show that if the hyperscalers continue on the current trajectory, the AI boom will become a story of one of the largest destructions of shareholder value in history. But there are two ways out of this corner."}],[{"start":null,"text":""}],[{"start":149.9,"text":"One is that the rise of AI will generate much more revenue for these companies than currently expected. Unfortunately, that comes with its own mathematical challenges. If the hyperscalers want to generate, say, a 10 per cent return on investment, they would have to find an additional $2tn to $5tn in revenue a year. A tall order for a group of companies that currently generates revenues of just $1.5tn per year."}],[{"start":177.15,"text":"The other option is that the planned investment in data centres, computer chips and other areas never materialises — maybe as equity investors turn more cautious on the sector or if debt funding for data centres becomes harder to get."}],[{"start":190.70000000000002,"text":"What then will happen when these companies announce cutbacks on some of their investment plans? The share prices of the largest companies on every continent from Nvidia to ASML, Samsung and TSMC are supported by these investment plans and the resulting demand."}],[{"start":206.70000000000002,"text":"And remember that US GDP growth currently is driven exclusively by rising tech spending. If these start to drop, the US economy will enter recession very quickly — even if tech investments decline only by a little bit, say 4 to 6 per cent, as happened after much smaller tech booms in the 1960s and during the 2009 recession."}],[{"start":227.00000000000003,"text":"Such a mild correction in investment spending is likely to still send stock markets in the US, UK and across Europe into a new bear market. A repeat of the tech crash in the early 2000s is a real risk with stock market drops of 50 per cent or more in the first year."}],[{"start":244.30000000000004,"text":"The question then becomes, when could we see such announcements by the hyperscalers? I think it is unlikely to happen in 2026. The efforts by OpenAI, Anthropic and others to keep the hype going at least until after their respective IPOs are likely to support the boom for now. But then what? The impossible maths of hyperscalers stays the same, but the marketing hype may fade. And eventually, reality will kick in. Probably not in 2026, but possibly in 2027 or 2028. After all, Greenspan talked about irrational exuberance in December 1996, but the bubble only burst three years later in 2000."}],[{"start":284.70000000000005,"text":"Viewed through this lens, the IPO of these AI companies is probably nothing more than a major transfer of investment risk from the current owners to retail investors, pension funds and others who are willing to buy the hype. "}],[{"start":302.8,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1779453068_3589.mp3"}