Key insights
- Data center stocks faced pressure after Elice, a South Korean cloud company, saw its shares plunge 30% on its market debut. This event, coupled with concerns about overcapacity and moderating AI revenues from companies like Anthropic and OpenAI, is fueling bearish sentiment for AI-dedicated computing power providers. HSBC analysts maintain a 'reduce' rating on CoreWeave, citing debt market woes and structural concerns.
In this article
Data center stocks are being squeezed on Tuesday following a lackluster initial public offering overnight from a South Korean cloud company that has been compared to U.S. neocloud CoreWeave.
Shares of Elice dropped 30% during their market debut overnight, as investors grow increasingly wary of artificial intelligence-dedicated computing power providers amid concerns about overcapacity.
The bearish sentiment spilled into U.S. tech equities before some names clambered back into the green. CoreWeave shares are up 1.6% after dropping 4% just after market open Tuesday; Nebius shed about 1% while Iren was down about 4% in late morning trading.
HSBC said in a Monday note that the troubles for the heavily levered CoreWeave are not going away.
“Debt market woes and moderating LLM revenues are exacerbating structural concerns for neoclouds,” HSBC analyst Abhishek Shukla wrote. Shukla has a reduce rating on CoreWeave with a price target of $39, implying 36.5% downside.
Elice’s IPO was small, having raised just $150 million, according to Reuters. Its market cap also equates to roughly $1.3 billion. The company provides a range of AI-related services, including cloud computing equipment, training libraries and application interfaces for businesses.
Despite its size, investors are likely looking at it as a a signal of investor sentiment toward the AI build-out.
Trouble for neoclouds, which some investors see as the weakest link in the emerging AI value chain, started earlier this month with less-than-stellar revenues reported by the private AI algorithm companies and pressure from higher interest rates.
Frontier model Anthropic showed a deceleration in quarterly revenues, with sales dropping from 140% growth in the second quarter to 45% in the third. OpenAI also disappointed, reporting revenue that was $20 billion shy of what analysts expected, analyst Stephanie Roth at Wolfe said in a Sunday note.
Profitability concerns for AI algorithm companies were the focus of a report last week from advisory firm Gartner predicting that 80% of neoclouds could cease independent operations by 2030.
“Our research indicates that enterprise and consumer AI revenue will be insufficient to support profitable returns on all planned AI capital expenditures” between 2027 and 2030, Will Sommer, senior director analyst at Gartner, said.
Gartner stood by its analysis in a Monday interview with CNBC.
“At the moment, we have more compute capacity than we have demand, and we have a lot of the tokens and processing power being given away in freemium solutions,” Gartner chief forecaster John Lovelock told CNBC. “Everyone in the AI race is running for a last-man standing position. Pretty much every market that Gartner is tracking in our AI race is likely to collapse down to a duopoly.”