Credit markets are increasingly targeting CoreWeave as investors worry that companies at the center of the artificial intelligence boom will fail to generate the cash flows necessary to service their massive liabilities. The price of a five-year bond issued by the neocloud provider has dropped significantly since July, with its yield surging to levels typically associated with financial distress. Credit-default swaps indicate that the likelihood of the firm struggling with its debt over the next five years has risen sharply, spilling over into a wider tech stock market wobble.
CoreWeave is not alone, as yields on bonds from rival neocloud providers like Nebius and TeraWulf are also climbing. Out of ten publicly listed neoclouds, only two are currently profitable, with their combined borrowings tripling their annual revenues. Across the broader technology landscape, Oracle, SoftBank, SpaceX, and the major hyperscalers including Alphabet, Amazon, Microsoft, and Meta have drastically scaled up their borrowing this year to fund infrastructure and data-center development.
When including chipmakers, semiconductor supply chain firms, and electricity utilities upgrading their power capacity, total borrowing related to the AI boom has reached nearly eight hundred fifty billion dollars for the year. Collectively, the balance sheets of these organizations now carry more than three trillion dollars in debt. JPMorgan Chase estimates that capital spending to support the technology will reach five and a half trillion dollars by 2030, with three-quarters of that figure expected to be funded through debt.
To manage this escalating financial burden, borrowers are utilizing creative financing strategies and tapping international credit markets. Roughly thirty percent of bonds issued this year by the five largest cloud providers have been denominated in non-dollar currencies such as euros, pounds, and yen. At the same time, special-purpose vehicles are becoming commonplace to keep obligations off corporate balance sheets, exemplified by joint ventures between major cloud operators and asset-management firms to fund large-scale data-center campuses.
Additional complexity stems from chip-backed financing, where deployed silicon serves as collateral for hardware purchases through separate legal entities. Prominent model-makers and cloud providers are increasingly exploring these arrangements to secure semiconductors. Furthermore, firms carry hundreds of billions in lease liabilities, trillions in future lease commitments, and extensive supply purchase obligations, leaving ratings agencies and investors to navigate an increasingly opaque and interconnected web of risk.