Gartner Reaffirms AI Bubble Thesis: “80% of Neoclouds Cannot Survive Independently”

[Financial News]  

Gartner, a Wall Street research and advisory firm, reaffirmed its longstanding position that investment in artificial intelligence (AI) is a bubble. It warned that, without sufficient demand, 80% of neoclouds would be unable to survive independently by 2030.
Gartner’s bearish analysis triggered a sell-off in AI-themed stocks last week.
Gartner Vice President John Lovelock, its chief forecaster and an analyst, warned in an interview with CNBC on the 19th (local time) that the $3.5 trillion in AI investment (about 4,665 trillion won) could result in massive losses. He also predicted that mergers and acquisitions among neocloud companies would be inevitable because the timing was wrong.
“We currently have more computing capacity than demand, and we’re giving away enormous amounts of tokens and computing capacity under the freemium model,” Vice President Lovelock said. He added, “Everyone in the AI race is fighting fiercely to be the last one standing,” and said, “In nearly every market Gartner currently tracks, the AI race will ultimately take the form of a duopoly.”
He predicted that the duopoly would be built around the commercial frontier models from Anthropic and OpenAI.
Lovelock said the computing capacity neoclouds have raced to build would be difficult to sustain independently on a profitable basis, and predicted that mergers among the companies would be inevitable.
Last week, Gartner predicted in an analysis note that 80% of neocloud companies would be unable to survive independently on a profitable basis and would end up merging. The computing capacity of these AI cloud companies would still be needed, it said, but continued massive losses would make it impossible for them to stand alone.
The report sent leading neocloud provider CoreWeave tumbling 25% last week, while Nebius and IREN plunged 25% and 16%, respectively. Those declines far outpaced the 2% drop in the S&P 500, a market benchmark, and the 4% drop in the tech-heavy Nasdaq Composite.
Lovelock also said AI would not be a standalone industry but would instead play the role of salt added to existing industries. As AI is applied across all industries, it will naturally become embedded in the market, he explained.
Meanwhile, in an analysis note on the 14th, Gartner said, “AI revenue will not be sufficient to support the profitability of all AI capital expenditures planned for 2027–2030,” triggering a sell-off in AI stocks.
Gartner noted in particular that Anthropic’s latest quarterly revenue of $16.8 billion was “a drop in the bucket” compared with the more than $5 trillion in massive investment hyperscalers have planned through 2030. The term hyperscalers refers to companies such as Amazon.com, Inc., Google and Microsoft (MS), which provide cloud services to third parties through large-scale data centers.

[email protected] Song Kyung-jae Reporter