The artificial intelligence (AI) party remains in full swing, with tens of billions flowing into infrastructure, startups and talent. High-profile commitments this year include OpenAI, SoftBank and Oracle pledging $500 billion for AI supercomputers, an OpenAI–Nvidia $100 billion fund for advanced chips, and heavier investments from China’s Alibaba and Tencent as Beijing pursues leadership in AI by 2030. Since ChatGPT’s November 2022 debut, AI-related stocks have added an estimated $17.5 trillion in market value, driving roughly 75% of the S&P 500’s gains and lifting firms such as Nvidia and Microsoft to record valuations.
Corporations are hesitant over AI adoption
Signs of a hangover are emerging. Corporate AI usage is slipping, spending is tightening, and machine-learning hype has massively outpaced profits. Many economists argue that usage concerns, barely three years after AI went mainstream, undermine the narrative that AI will revolutionize business by automating repetitive tasks and improving forecasting.
“The vast bet on AI infrastructure assumes surging usage, yet multiple US surveys show adoption has actually declined since the summer,” Carl‑Benedikt Frey, professor of AI & work at the University of Oxford, told DW. “Unless new, durable use cases emerge quickly, something will give — and the bubble could burst.”
The US Census Bureau’s fortnightly survey of 1.2 million US firms found AI-tool usage at companies with more than 250 employees fell from nearly 14% in June to under 12% in August. AI’s practical weaknesses — hallucinations that produce plausible but false information, inconsistent reliability, and autonomous agents that complete tasks successfully only about a third of the time — compound adoption challenges. “Unlike an intern who learns on the job, today’s pretrained [AI] systems don’t improve through experience. We need continual learning and models that adapt to changing circumstances,” Frey said.
Unsustainable capital burn
As the gap widens between expectations and commercial reality, investor enthusiasm is cooling. In Q3, venture-capital deals with private AI firms fell 22% quarter-on-quarter to 1,295, although funding remained above $45 billion for the fourth consecutive quarter, according to CB Insights.
“What perturbs me is the scale of the money being invested compared to the amount of revenue flowing from AI,” economist Stuart Mills, a senior fellow at the London School of Economics, told DW. Market leader OpenAI, backed by Microsoft, reported $3.7 billion in revenue last year versus $8–9 billion in operating expenses. The company says it expects $13 billion this year, but was estimated by The Information to burn through $129 billion before 2029.
Mills argues that generative AI products like Grok and ChatGPT are “charging far less than they need to make a profit” and should raise subscription prices. Julien Garran of MacroStrategy Partnership has been even starker: he estimates a misallocation of capital equal to 65% of US GDP — four times the housing buildup before the 2008 crisis and 17 times the dot‑com bust — driven by huge AI investment despite scant evidence of sustainable returns.
Investors increasingly cautious
Big Tech earnings have produced cautious optimism but also fresh doubts. Palantir’s Q3 revenue rose 63% year‑over‑year, yet its stock fell up to 7% on the report. AMD and Meta reported strong AI-related results but still faced market skepticism about sustainability. This disconnect between soaring valuations and shaky fundamentals worries analysts who say AI is not yet penetrating high up the value chain to generate clear value.
Nvidia’s upcoming earnings are widely watched as a test of the AI boom’s longevity. In Q2, Nvidia’s data-center sales accounted for 88% of revenue, which reached a record $46.7 billion; the company guided Q3 toward $54 billion.
When will the bubble pop?
“With the exception of Nvidia, which is selling shovels in a gold rush, most generative AI companies are both wildly overvalued and wildly overhyped,” Gary Marcus, Emeritus Professor at NYU, told DW. “My guess is that it will all fall apart, possibly soon. The fundamentals, technical and economic, make no sense.”
Garran believes rapid progress in large language models is slowing not due to technical limits but because the economics no longer stack up: training costs are skyrocketing while incremental improvements have diminished.
A more moderate view comes from Sarah Hoffman, director of AI Thought Leadership at AlphaSense, who predicts a “market correction” rather than a cataclysmic burst. After extended hype, enterprise investment will likely become more discerning, shifting “from big promises to clear proof of impact,” she said. “More companies will begin formally tracking AI ROI to ensure projects deliver measurable returns.”
Edited by: Uwe Hessler
