Elon Musk warns: Hundreds of AI companies in danger of disappearing
Elon Musk warns that hundreds of AI companies could disappear. Analysts expect a major remediation and Nvidia is also at risk.
The explosive growth of artificial intelligence (AI) has spawned a new multi-billion dollar industry. Companies are investing huge sums in data centers and powerful chips to meet the rising demand for computing power. But according to Elon Musk, a large part of this young sector is in danger of going under. There is also a growing expectation within the industry itself that extensive consolidation is inevitable.
Musk's warning focuses on so-called neoclouds. These are specialized companies that purchase powerful AI chips, build or rent data centers, and then lease computing power to technology companies. Due to the enormous interest in AI, the number of providers has grown significantly in recent years.
Musk predicted on Sunday on his platform X that more than 200 of these companies may disappear or be acquired. The main vulnerability is their dependence on the demand for AI computing power, while their investments are largely financed by debt.
Large technology companies can often also use their infrastructure for other cloud services. This is different for specialised providers. When demand for AI capacity is disappointing, they are left with costly data centers, rapidly aging chips, and bulky loans.
Chip manufacturer Nvidia is also at risk. The company has invested more than $7 billion directly in AI computing power providers. In addition, the most recent quarterly report shows that Nvidia has up to $164.5 billion in guarantees and liabilities related to data centers in which its chips are used.
The shares of CoreWeave, Nebius and Iren fell Monday morning after Musk's statements. However, by the end of the trading day, all three companies had made up for their losses.
The concerns are shared by various research agencies. Gartner expects that 80 percent of independent neocloud companies will no longer be self-employed by 2030. This does not necessarily mean that all these companies will go bankrupt. Acquisitions and mergers can also cause companies to lose their independence.
In addition, Bernstein analysts estimate that only 35 percent of the announced data center capacity in North America is actually being built, Bloomberg reported.
A major obstacle is the availability of electricity. New data centers require huge amounts of power, while connection to the grid is becoming increasingly difficult in many countries. Local opposition to the construction of data centres is also causing delays.
Raimund Riedl, financial director of AI infrastructure company Volta, also expects a radical change. At an Nvidia meeting in Berlin, he told Fortune that the industry is on the verge of a huge wave of consolidation.
Volta seeks to differentiate itself by attracting institutional investors. The company previously raised $300 million from investors, including Andreessen Horowitz and Nvidia. In addition, it has a $10 billion contract, reportedly with Anthropic. That contract helped obtain a $5 billion financing agreement.
According to Paul Miller, analyst at Forrester, the expected consolidation is a logical consequence of the huge interest in artificial intelligence. Many new companies are trying to take advantage of the growing market, but not everyone has enough capital, customers and infrastructure to survive.
Miller expects the first acquisitions and mergers to become visible within a year, after which consolidation can accelerate further.
According to him, that does not mean that the growth of AI data centers is over. The demand for computing power remains high, but many companies have promised capacity that they may never be able to deliver.
For established providers, this can create opportunities. Companies that already have working data centers, powerful chips and existing customers can benefit when smaller competitors disappear.
The expected remediation therefore mainly revolves around the question of which companies can actually realise their ambitious plans. Not only the demand for AI, but also access to electricity, financing and reliable infrastructure will determine who will survive.