Copyright cityam
Federal reserve chair Jerome Powell recently poured cold water on mounting fears that the AI boom is heading for a spectacular bust, arguing that the current surge in investment bears little resemblance to the dot-com frenzy that ended in tears two decades ago. In a press conference following the Fed’s latest rate decision this week, Powell stated that today’s AI leaders are not riding a wave of speculation, but rather building real, profitable businesses, and in doing so, powering a crucial leg of America’s economic growth. “This is different in the sense that these companies [today] that are so highly valued actually have earnings and stuff like that,” he said. “If you go back to the nineties and the dot-com, these were ideas rather than companies. So there was a clear bubble there.” His words come in the wake of America’s central bank cutting its rates by 25 basis points, a largely expected move, while investors digested another round of earnings from Big Tech. Apple, Amazon, Meta and Google all pledged to ramp up AI infrastructure spending, each brushing off concerns that enthusiasm could overheat the market. Trillion-dollar question The root of mounting concerns over a potential bubble lies in the scale of investment. Alphabet, Microsoft, Amazon, and Meta are on track to invest nearly $370bn in AI-related capital expenditures this year alone, primarily to fund new data centres and semiconductors. Google’s chief financial officer, Anat Ashkenazi, told investors this week that the company’s AI budget had risen to as much as $93bn, a huge uptick from previous forecasts. Meanwhile,Meta expects to spend at the top end of its $66–72bn range and warned that next year’s AI investment will be “notably larger.” Microsoft, which is partnering with OpenAI, plans to increase its spending beyond $80bn in 2026 as it builds new AI infrastructure across the US. Those numbers make the early internet boom look modest. In the late 1990s, companies invested billions in fibre-optic cables that would later sit idle, as analysts at the time called them “dark fibres”. But today, as Atreides Management’s Gavin Baker argues, “there are no dark GPUs.” The chips powering the AI revolution are being used as fast as they can be made. Growth engine Powell’s point was that the AI surge is being driven not by easy money, but by long-term confidence in its transformative potential. “I don’t think interest rates are an important part of the AI or data centre story,” he said. “It’s clearly one of the big sources of growth in the economy.” Similarly, Goldman Sachs’ chief US economist, Joseph Briggs, estimated that the productivity gains from AI could be worth between $8tn and $19tn in present value for the US economy. JPMorgan, meanwhile, forecasts that AI-related infrastructure spending could boost GDP growth by around 0.2 percentage points over the next year. Powell’s comments landed just hours after several of the world’s largest tech firms unveiled fresh spending plans that cement their role at the centre of the economy. Google, Meta and Microsoft said they would spend more than previously expected on AI infrastructure this year, even as the Fed acknowledged a ‘softening’ labour market. Not everyone is convinced But there are still objectors. Bill Gates, for one, seems sure that the world is caught up in this so-called “AI bubble”, comparing the current rush of capital to that of the dotcom era. “Some companies succeeded, but a lot of the companies were kind of me-too, burning capital”, Gates said. “Absolutely, there are a ton of these investments that will be dead ends.” This was also echoed by David Einhorn, founder of hedge fund Greenlight Capital, who told Bloomberg that “the numbers being thrown around are so extreme that it’s really hard to understand them. There’s a reasonable chance that a tremendous amount of capital destruction is going to come through this cycle.” Even some of the tech industry’s leading figures have voiced similar concerns, with Meta boss Mark Zuckerberg admitting that the current wave of AI spend “could create a bubble”, despite arguing that the risk was worth taking. “It’s the right strategy to aggressively front-load building capacity,” he said on Meta’s earnings call. “In the worst case, we would just slowly build new infrastructure while we grow into what we’ve built.” ‘No dark GPUs’ But others insist that comparing AI to past bubbles missed the point entirely. Nvidia chief executive Jensen Huang whose company just hit a market cap of over $5tn, flatly rejected these concerns. “I don’t believe we’re in an AI bubble,” Huang said at Nvidia’s GTC DC event. “All of these different AI models we’re using, we’re using plenty of services and paying happily to do it. Customers are willing to pay for powerful AI models, and the market remains healthy.” And the numbers seem to back the tech titan, with Nvidia’s latest chips expected to generate half a trillion dollars in revenue. The company’s processors are in such high demand that, in the words of Gavin Baker, managing partner of Atreides Management, GPUs are “melting” from overuse. Global AI-related expenditures are projected to reach $1.5tn by next year and exceed $2tn in 2026, according to research firm Gartner, nearly two per cent of global GDP. Those figures reflect not speculative bets, but tangible corporate spending, which is already reshaping the business landscape. A long-term wager Powell argues that this is the key difference between ambition and seeing returns. The dotcom bubble was driven by unproven startups chasing clicks, whereas today’s AI race is led by profitable heavyweights deploying capital at a scale unseen since the postwar industrial build-out. “Consumer spending is much bigger than that and has been growing, and has defied a lot of the negative forecasts,” Powell said, underscoring that the AI boom is additive, not the sole driver of the economy. Yet even he acknowledges the irony that the same technology that promises to boost productivity could also reshape, and potentially shrink, parts of the labour market. Layoffs at major firms are increasingly attributed to automation, while the productivity gains may take years to show up in official data. For now, though, investors appear to agree with Powell and Huang that this is not 1999. The AI trade remains intact, as Gene Munster of Deepwater Asset Management put it this week, and “these companies are continuing to talk about spending and investing much more than we thought three, six months ago.” It seems that for all the warnings of overreach, the bets are still being placed, and the chips, quite literally, are still hot.
