The great AI sell off - a pause for thought, or a panic?

A sell-off in AI-related stocks this week has raised questions about whether the market is experiencing a healthy pause or a panic, as investors reassess valuations amid massive capital spending by tech giants on AI infrastructure. The downturn was triggered by concerns about AI demand and technical market pressures, though analysts are divided on whether it reflects profit-taking or deeper anxiety about returns failing to match hype.

The great AI sell off - a pause for thought, or a panic? Rockets and robots: the rise of financial fabulism - Bookmark - CommentsGo to comments Elon Musk /topic/elon-musk is on the skids. The paper value of his shares in SpaceX /topic/spacex and Tesla have tumbled in recent days, leaving the entrepreneur /topic/entrepreneur — an eccentric genius or a danger to all of us, depending on your point of view — significantly poorer. He is down to his last $957bn. Perhaps we should have a whip round /news/business/elon-musk-net-worth-trillionaire-fortune-spacex-tesla-b3001927.html . The falls in SpaceX and Tesla might not mean much in isolation. Shares in high-growth technology companies /topic/technology-companies have always been prone to bouts of euphoria followed by equally dramatic corrections. But there is no escaping the wider trend that has rattled markets /topic/markets this week. After months of seemingly relentless gains, the AI /topic/ai trade /topic/trade suddenly hit turbulence. The era of SpaceX, Nvidia and the hyperscalers has been built on what some critics describe as financial fabulism: the idea that the future will be bright thanks to brilliant inventions that, erm, don’t exist yet. Investors have spent the past two years buying into a vision of an AI-powered future that promises to transform productivity, reshape industries and generate vast profits. The trouble is that much of that future remains theoretical. What is not theoretical is the spending. Alphabet /topic/alphabet , Amazon, Meta and Microsoft are expected to spend as much as $720bn this year, much of it on AI infrastructure /topic/infrastructure and data centres. These are sums that would once have been associated with governments rather than corporations. The race to dominate artificial intelligence has become one of the largest capital expenditure programmes in modern business history. Until recently, investors were happy to fund it. Now they appear to be asking an awkward question: what if the returns fail to match the hype? The sell-off that swept through technology stocks this week was triggered by a mix of factors, from concerns about AI-related demand to technical market pressures. But beneath the immediate catalysts lies a deeper anxiety about valuations. Matthew Ryan, head of market strategy at Ebury, argues that the episode highlights just how nervous investors have become. “Futures have stabilised this morning, though this latest episode once again shows just how fragile risk sentiment can become when AI valuation fears are anywhere near the surface.” That fragility is striking because there has been little genuinely negative news. The AI story itself has not materially changed. Demand for computing power remains immense. Companies are still scrambling to build data centres. Consumers continue to embrace AI tools at remarkable speed. Lale Akoner, global market strategist at eToro, sees the move less as panic than a reality check. “This looks more like profit-taking and rotation than outright panic,” she says. “The AI trade has delivered extraordinary gains, so it is unsurprising to see investors locking in returns after such a powerful run.” The distinction, she argues, is between the long-term fundamentals and short-term positioning. Demand for AI infrastructure remains robust, but when investors crowd into the same trade, even relatively minor developments can trigger sharp swings. That appears to be what happened this week. Concerns surrounding South Korean memory-chip giant SK Hynix, combined with leveraged exchange-traded funds that can magnify moves in popular technology stocks, helped create a feedback loop of selling pressure. Lindsay James, investment strategist at Quilter, notes that the immediate triggers alone do not fully explain the scale of the market reaction. Instead, she points to a market that may have become excessively exposed to a handful of AI winners. “The volatility may be due to both the leveraged single stock ETF structures in the Korean market,” she says, warning that these products can become forced sellers when stocks decline, amplifying losses across the sector. The result is a reminder that financial manias are rarely driven solely by fundamentals. They are also driven by flows of money, investor psychology and the assumption that someone else will always be willing to pay a higher price tomorrow. Yet this is not the dotcom bubble all over again — at least not yet. Unlike many of the companies that fuelled the excesses of the late 1990s, today's AI giants are enormously profitable businesses. They generate real cash, dominate their markets and possess technological capabilities that are genuinely transformative. The question is not whether artificial intelligence will change the economy. It almost certainly will. The question is whether it can do so quickly enough to justify the staggering sums currently being invested. That challenge becomes even more significant as companies increasingly turn to investors and bondholders to finance the AI arms race. Alphabet has raised fresh capital through stock sales. Amazon has tapped debt markets on a vast scale. Even Musk's SpaceX is preparing to spend heavily on AI ambitions alongside its space programme. And there is no sign that the spending spree is ending. OpenAI, which is reportedly considering a future public listing at a valuation of around $730bn, could soon become the most dramatic test yet of investor appetite for AI. As Minmo Gahng, assistant professor of finance at Cornell University notes, OpenAI enjoys something few private companies possess: mass recognition. “That kind of household-name recognition could generate substantial retail demand and support a richer valuation than fundamentals alone would justify.” Those last four words may be the most important in the entire AI debate. For now, markets are not questioning whether artificial intelligence matters. They are questioning whether investors have got ahead of themselves. The AI revolution may yet transform the world. But this week has been a reminder that revolutions are expensive — and sooner or later, somebody has to show the receipts. Join our commenting forum Join thought-provoking conversations, follow other Independent readers and see their replies Comments comments-area