Musk and other Silicon Valley figures talk down new cloud providers, with idle compute and tightening financing in focus

Article body

Recently, when replying to well-known Silicon Valley venture capitalist Chamath Palihapitiya on social media, Elon Musk said that over the past few weeks, the debt market has almost cut off financing channels for new cloud providers whose revenue depends entirely on AI data centers.

In his view, this means that more than 200 neoclouds (new-type cloud service providers) will face guppy-level death and consolidation. "Some (neoclouds) can survive the difficulties, but very few," he said.

Palihapitiya pointed out that, as an investor with a large amount of developed land assets, he has observed problems at neocloud platforms such as organizational chaos and insufficient capitalization.

He stressed that signing a confidentiality agreement with a hyperscaler takes 3 months and also requires an 18-month hosting period. In 2027, there will be more stranded computing power and inventory computing resources than currently forecast; these resources have already been procured and paid for, but in practice face insufficient returns or sit idle and gather dust. Musk echoed this: "I think we will see a lot of computing power resources sitting idle in data centers that have not yet been fully built."

Recently, against the backdrop of a renewed tightening in the interest rate environment and the dual policy and financing uncertainties facing data center construction, discussion of the risk that neoclouds are "short of money" has been heating up.

Gartner, a globally renowned consulting firm, predicts that by 2030 more than 80% of neocloud service providers will be forced to terminate independent operations, with the core reason being that the industry's revenue gap will reach the trillions of dollars level. The firm estimates that AI data centers will need to generate $8.5 trillion in revenue by 2030 to cover costs, but actual revenue is expected to be only $3.6 trillion.

At the same time, pressure has already appeared in the AI infrastructure financing market. According to reports, CoreWeave has about $50 billion in debt, and the yield on some of its bonds once rose to around 15%; the company's capital expenditure this year is expected to reach as high as $39 billion, while analysts expect its 2026 revenue to be close to $13 billion and its loss to be about $2.13 billion.

CITIC Securities judges that the core market concern is not only the pressure that high leverage and high financing costs place on neocloud providers' cash flow and debt-servicing capacity, but also that related risks may spread to the AI industry chain. It believes that the market has overamplified this extreme scenario in the short term, and that the overall debt pressure of neoclouds is still a reflection of AI demand.

This coincides with Jensen Huang's recent remarks: at the GTC conference held in Berlin, he stressed that market demand for new-type cloud companies is "off the charts," but a shortage of funds is hindering growth. He said: "What is missing in the middle is the working capital needed to drive growth—because there is huge demand in every area of artificial intelligence, so there is a shortage of funds."

The above institution further pointed out that although AI capital expenditure has brought high debt problems to neoclouds and related enterprises, considering various financing methods such as public markets and GPU collateral, it believes that the AI debt problem in 2027 will remain controllable, but it is necessary to observe changes in AI at the model capability and commercialization levels after entering 2027H2, as well as subsequent trends such as rental prices.

Huatai Securities believes that, in the environment of generally high interest rates in overseas developed countries and considering event-driven influences such as midterm elections, AI financing risks may continue to ferment. This may further increase the construction cost of data centers and drag down the pace of implementation. The rise in construction costs and financing costs may intensify the internal differentiation among cloud providers; hyperscalers with more existing data centers and higher credit ratings may be more advantaged, and attention should be paid to whether the third-quarter reports can bring new catalysts.

(Source: Cailian Press) Body middle Embedded advertisement