image copyrightReuters
A previously little-known area of the artificial intelligence (AI) ecosystem is now in the spotlight, and investors have mixed views on its prospects.
Company names like CoreWeave and Nebius are familiar to tech circles in San Francisco and Silicon Valley.
These data center companies, known as "neocloud" companies, operate a large number of data centers filled with expensive computer equipment and lease computing power to large artificial intelligence developers such as Anthropic, OpenAI and Meta.
Today, they are one of the most hotly debated topics in Silicon Valley.
Many people have questioned whether the financial health of neocloud companies is sound, and whether their recent setbacks reflect the current state of the artificial intelligence economy.
Analysts at business and technology research organization Gartner predict that by 2030, 80% of new cloud companies will "stop operating independently." On the other hand, experts from Forrester Research expect that at least three new cloud companies will be acquired in the short term because their "financial conditions are becoming unstable."
U.S. Treasury Secretary Scott Bessent tried to downplay outside concerns about new cloud companies last Thursday (October 29). In an interview with the US Consumer News and Business Channel (CNBC), he said that these companies are only "a small part of the overall ecosystem."
debt worries
New cloud companies typically raise capital through long-term leases with large technology customers that need quick access to computing power.
End of Hot Reading
Cory Johnson, an analyst at the initial public offering (IPO) research team "The Quiet Period", explained that these companies can then use the relevant contracts and the purchased artificial intelligence chips as collateral to borrow money.
He pointed out that debt is quite common among new cloud companies because the related hardware equipment requires up-front costs.
Industry leader CoreWeave, for example, is much smaller than some of the big tech companies that use its services.
But Johnson believes that lenders sometimes give more weight to the creditworthiness of CoreWeave customers such as Microsoft and Meta than to CoreWeave's own financial health when granting loans.
"This helps explain how CoreWeave accumulated US$35.1 billion (233.7 billion yuan; NT$1.15 trillion) of debt," he said.
Some uneasy investors appear to be changing their views.
CoreWeave shares have lost a quarter of their value over the past month, and that's not an isolated case.
Other new cloud companies, including Nebius, IREN and Applied Digital, have seen their share prices fall by similar or even greater margins during the same period.
Bond investors also appear to be spooked and are beginning to demand higher returns from several companies in the industry.
CoreWeave CEO Mike Intrator said the market's current reaction is "irrational" and is confident in the company's business model.
Some of the world’s most influential companies are lining up to use CoreWeave’s infrastructure and other AI services.
Origin of cryptocurrency
Some of the new cloud companies emerge from the volatile cryptocurrency mining industry, which produces cryptocurrencies such as Bitcoin.
The infrastructure required for mining has a lot in common with the equipment required for the development of artificial intelligence.
CoreWeave was founded in 2017 and initially focused on Ethereum mining.
IREN, Core Scientific and Applied Digital also started out as cryptocurrency mining businesses.
When OpenAI launched ChatGPT in 2022, setting off the artificial intelligence craze, these companies' existing infrastructure gave them an advantage quickly.
Nebius is a little more special. The company was formerly known as Yandex NV, the parent company of Yandex, known as the "Russian version of Google", and will complete its spin-off and name change in 2024.
Today, the success or failure of these companies depends largely on large AI developers and so-called hyperscalers. The latter is a larger data center operator, and new cloud companies often regard it as an important customer.
"There's a lot of money pouring into the system right now," said Melissa Otto, director of research at S&P Global's Visible Alpha.
"This is an arms race."
And new cloud companies are one of the beneficiaries.
However, Charlie Dai, vice president and chief analyst at Forrester, said the price of this business model is that companies will face "higher financial risks" should the environment change.
That includes any changes to spending plans by big tech customers.
Just in the past week, when Microsoft and Google predicted that the scale of future artificial intelligence investment would be lower than Wall Street analysts expected, the stock prices of many new cloud companies fell in response.
In addition, the market is also paying close attention to the development of Anthropic and OpenAI. The companies expect to launch a public offering of shares in the coming months.
Bradley Tusk, founder of the venture capital firm Tusk Ventures, said that if artificial intelligence turns out to be just a bubble and the bubble bursts: "New cloud companies will definitely be affected."
"Guppy-level mass death"
The future of new cloud companies also depends on their ability to continue borrowing, and when interest rates rise, financing costs will also increase.
Elon Musk recently posted on social platforms that this challenge may cause more than 200 new cloud companies to face "guppy-level death and mergers," with the smallest companies among the most vulnerable.
"Some companies will be able to survive, but not many," he added.
For now, analysts at Yorkville Ives said the market's demand for computing power still exceeds supply, and new cloud companies are filling important gaps that large data center operators cannot fill.
However, in the eyes of some people, the recent fluctuations in the new cloud industry are worrying.
Microsoft recently canceled a cooperation agreement with Nscale, a new British cloud company, due to power supply issues, and the power supply restrictions may last until 2030.
Australian company Firmus has just shelved its IPO plans.
The developments come amid a growing public backlash against data centers and questions about how much real value artificial intelligence can bring to worker productivity.
"There is real reason to be concerned," Johnson said.
This article was originally written in English. We used artificial intelligence to assist in translation and was reviewed by BBC journalists before publication. Learn more about how we use artificial intelligence.
Explore more
More short films
More short films end
