Gartner is maintaining its forecast that $3.5 trillion in AI investment faces impairment risks because current profitability does not match the scale of capital expenditure. John Lovelock, the firm's chief forecaster, stated in an interview with CNBC that returns are not yet backing up the full scale of investment, a view that helped trigger significant volatility in AI-related equities last week.
Lovelock explained that the market currently possesses more compute capacity than demand, with a substantial portion of processing power being distributed through freemium solutions. He described the competitive environment as a race for a last-man standing position, predicting that nearly every tracked AI market will collapse into a duopoly. For the main commercial frontier models, that duopoly consists of Anthropic and OpenAI.
The mismatch between the massive asset base being built by neoclouds and their current cash flow is central to Gartner's analysis. While Lovelock clarified that neoclouds will not go out of business en masse, he warned that consolidation through takeovers and mergers is likely. Gartner's Wednesday report predicted that 80% of neoclouds, defined as AI-dedicated computing power providers, could cease functioning as independent companies by 2030.
The impact on markets was immediate. Neocloud stocks fell sharply last week, pulling down broader indices. CoreWeave and Nebius each dropped 25%, while IREN fell 16%. The S&P 500 finished down 2% and the Nasdaq Composite fell 4% for the week ending Oct. 16, as investors rotated away from technology stocks. The Dow Jones Industrial Average finished up 0.7%. On Monday, some of these stocks rebounded, with CoreWeave up about 2%, Nebius up 3.3%, and IREN up 5.6% in midday trading.
Stephanie Wolfe at Roth noted in a client report that the Gartner findings likely contributed to the market stress. She highlighted that Gartner warned more than $3.5 trillion in AI infrastructure spending could be exposed to impairment by 2030, with neocloud providers particularly vulnerable. Gartner researchers concluded that incoming AI revenue will not cover the costs of all planned capital spending between 2027 and 2030.
This finding coincides with decelerating revenue growth at frontier model companies. Anthropic's growth rate fell from 140% in the second quarter to 45% in the third quarter, reaching a quarterly revenue level of $16.8 billion. This figure is small relative to the more than $5 trillion that hyperscalers are expected to invest through 2030. Additionally, OpenAI's annual revenue expectations missed by $20 billion the previous week, according to Roth.
Lovelock characterized AI as a technology rather than a distinct market, comparing it to table salt sprinkled into existing operations rather than a new field for differentiated products. Other analysts have echoed these concerns. Chris Wood, Jefferies' global head of equity strategy, has warned about capital destruction in AI since September. Paul Tudor Jones told CNBC last week that macroeconomic factors would lead to a challenging year for the stock market in 2027, citing a negative flow picture driven by buybacks and M&A versus IPOs and secondary offerings.