The artificial intelligence boom shows no sign of slowing, with global spending on AI infrastructure set to surpass $600 billion this year. Yet beneath the headline figures, a growing number of investors are raising urgent questions about who, ultimately, will profit — and who may not survive.
A Flood of Capital Into AI
The scale of AI investment is almost without precedent. According to analysts at Natixis, the so-called Magnificent Seven technology giants — which include Microsoft, Google parent Alphabet, and Amazon — are expected to spend around $600 billion on data centres, chips, and AI infrastructure in 2026 alone. S&P Global puts that figure even higher, at approximately $635 billion. Credit markets, too, have shown a strong appetite for technology bonds, with large firms attracting substantial investor demand for recent debt placements.
The confidence stems from a belief that AI will reshape virtually every sector of the global economy. Since the launch of ChatGPT, roughly $10 trillion has been added to global equity market capitalisation, a staggering figure that reflects how deeply investors have embraced the AI narrative. Yet that enthusiasm is increasingly uneven.
Software Stocks Under Pressure
While AI chipmakers and infrastructure providers continue to attract capital, a very different story is unfolding in the software sector. Shares in some of the world’s most valuable software companies have tumbled sharply this year, as investors grow anxious that AI could render their core business models obsolete.
Salesforce has fallen around 21% this year, ServiceNow by 26%, Adobe by 22%, and Intuit by as much as 37%. An exchange-traded fund tracking the US software industry has dropped more than 22%. The concern is straightforward: if AI agents can perform tasks that businesses previously relied on dedicated software to complete — from payroll to document management — then demand for traditional software licences could collapse.
Nick Evans, a fund manager at Polar Capital whose global technology fund outperformed 99% of its peers over the past year, has been blunt in his assessment. “We believe application software is facing an existential threat from AI,” he told Bloomberg. He compared the current moment to the devastation wrought upon print media by the internet in the early 2000s, predicting that only a handful of software companies would successfully navigate the transition.
“Existential Question Marks”
What makes the current software sell-off particularly striking, analysts say, is that it is not being driven by overvaluation or investor exuberance — the usual culprits in tech downturns. Instead, it reflects what one investment director at GAM Investments described as “existential question marks” surrounding a business model that had previously commanded a significant valuation premium.
Arthur Mensch, chief executive of the French AI laboratory Mistral, has suggested that more than half of the enterprise software currently in use could potentially be replaced by AI. Yet some analysts urge caution about the scale of the panic. Michael Field, chief equity strategist at Morningstar, believes investor anxieties may be “overblown,” though he concedes a full recovery in software stocks is unlikely in the near term.
Resilience in Niche Areas
Not all software firms are equally exposed. Companies providing mission-critical systems or operating in highly regulated industries — where deeply embedded workflows and proprietary datasets create significant barriers to entry — are considered more resilient to disruption. SAP, which offers complex enterprise solutions, has been singled out as one firm better placed to weather the storm.
ServiceNow offers a telling example of the volatility at play. Earlier fears that AI agents would cannibalise its platform triggered a sharp sell-off, with its share price falling around 33%. Yet the company subsequently reported 21% subscription revenue growth and secured dozens of major enterprise AI deals, suggesting that, for some firms, AI may yet prove an opportunity rather than an executioner.
For investors, the challenge is distinguishing between the two.
