Data center stocks faced pressure on Tuesday, prompted by a disappointing initial public offering from South Korean cloud company Elice, which saw its shares plummet 30% at market debut. Investors are increasingly cautious about companies providing artificial intelligence-dedicated computing power due to concerns regarding overcapacity.

Read More

This negative sentiment affected U.S. tech equities, although some stocks recovered slightly later. CoreWeave's shares increased by 1.6% after an initial drop of 4% at the open, while Nebius and Iren saw declines of about 1% and 4%, respectively, in late morning trading.

HSBC analyst Abhishek Shukla noted in a report on Monday that challenges facing heavily leveraged firms like CoreWeave persist. He pointed out ongoing issues in the debt market and declining revenues from large language models (LLMs) as contributors to structural concerns in the neocloud sector. Shukla has assigned a 'reduce' rating to CoreWeave, with a price target of $39, indicating a potential downside of 36.5%.

Elice's IPO was modest, raising $150 million and giving the company a market capitalization of approximately $1.3 billion. Elice offers various AI-related services, including cloud computing solutions and training libraries.

The struggles facing neocloud companies have been compounded by recent earnings reports illustrating slowing revenues from private AI algorithm firms, alongside rising interest rates. For instance, Anthropic's revenue growth fell sharply from 140% in the second quarter to 45% in the third, while OpenAI reported revenues that were $20 billion below analyst expectations, according to Wolfe analyst Stephanie Roth.

Concerns about profitability in the AI sector were further emphasized in a report from advisory firm Gartner, which predicted that 80% of neoclouds could become unviable by 2030. Senior director analyst Will Sommer indicated that projected AI revenues would not be sufficient to justify the planned capital expenditures.

Gartner's chief forecaster John Lovelock reiterated this outlook in an interview with CNBC, highlighting a mismatch between compute capacity and demand, as well as the prevalence of freemium solutions diluting market value. Lovelock anticipates that the AI sector may soon consolidate around a duopoly.