Rising borrowing costs are beginning to test the financing model behind the artificial intelligence infrastructure boom, with Goldman Sachs estimating that hyperscale cloud providers could spend $1.2 trillion on capital expenditures next year while generating about $1.1 trillion in operating cash flow. The gap would leave the sector increasingly dependent on debt markets already pressured by heavy U.S. government borrowing and elevated Treasury yields.
Goldman Sachs estimates that at least 35% of hyperscaler buildout spending will require bond financing, implying more than $400 billion of new debt issuance. That demand arrives as the 10-year U.S. Treasury yield has moved above 5.3%, raising the benchmark cost for long-term corporate borrowing across technology, data-center and semiconductor companies.
The bank’s warning centers on a potential mismatch between ambitious AI infrastructure plans and the capital available to fund them. If tighter credit conditions delay data-center construction, projected chip and equipment demand for 2027 and 2028 may not translate into actual purchase orders. Goldman described the risk as an “air pocket” for parts of the semiconductor supply chain.
CoreWeave’s debt signals growing scrutiny
Market concern became more visible on Oct. 14, when AI cloud provider CoreWeave fell more than 15% during the trading session and slid almost another 6% after hours. The company has become one of the market’s most closely watched examples of a heavily financed AI infrastructure buildout, combining large customer commitments with substantial funding requirements.
CoreWeave’s $2.75 billion unsecured bond, issued in April, traded below 80 cents on the dollar, according to market pricing cited in the supplied materials. That pushed its yield to maturity above 15%, a level that indicates traders are demanding a far greater return to accept the company’s credit risk.
Dow Jones market data showed CoreWeave’s five-year credit default swaps, which function as insurance-like contracts against default, rising to 881 basis points. Pricing in those contracts implied a five-year default probability above 50%, though such market-implied probabilities are not forecasts of a company’s actual failure rate.
Michael Intrator, CoreWeave’s chief executive officer, said in a memo that the company had $104.2 billion in undelivered orders and added more than $25 billion in new customer commitments during the third quarter. Those figures underscore the tension facing AI infrastructure providers: contracted demand may be substantial, but delivering it requires servers, networking equipment, power capacity and data centers before much of the related revenue is recognized.
Selling spreads through data-center and chip names
The pressure was not limited to CoreWeave. Oracle shares fell more than 7% on the day, while Dell declined more than 10%, according to the supplied market data. Applied Digital and IREN each dropped close to 10%.
Goldman Sachs’ technology, media and telecommunications sector tracking showed declines over the preceding five trading days of 8.5% for U.S. memory companies, 7.9% for AI data-center firms and 7.6% for semiconductor stocks. The S&P 500 fell 0.17% over the same period.
That divergence suggests markets are repricing the cost of financing AI capacity rather than simply reassessing demand for AI computing. Hardware suppliers can report strong order books while still facing delayed spending decisions from customers whose debt costs rise faster than expected.
ASML’s market reaction offered another example. The semiconductor equipment manufacturer reported third-quarter results and raised the lower end of its 2026 outlook, yet its shares fell as much as 3% after the release. Traders focused on the absence of higher forecasts for 2027 and beyond, when the financing burden from current data-center plans may become more visible.
Cash flow is turning negative at major platforms
Goldman Sachs’ equity strategy team cited mounting free-cash-flow pressure among the largest technology platforms funding AI capacity. Alphabet recorded its first single-quarter negative free cash flow in the company’s history in the second quarter, with a net outflow of $5.9 billion, according to the bank’s figures. Amazon reported negative free cash flow of $7.6 billion over the previous 12 months.
Negative free cash flow does not automatically signal financial distress for companies with large balance sheets and profitable core businesses. It does show how quickly spending on chips, data centers and supporting infrastructure can consume internally generated cash, increasing the appeal of bond markets and share issuance.
Goldman estimated that roughly $430 billion raised through seasoned U.S. equity offerings this year included 45% directed toward AI-related companies. Higher interest rates reduce the relative advantage of debt over equity, leaving companies with fewer painless options when they need to finance multiyear infrastructure projects while preserving cash for operations, acquisitions or share repurchases.
Treasury supply is competing for the same capital
The financing strain is being amplified by the U.S. government’s own borrowing requirements. The supplied materials cite a fiscal 2026 deficit of $1.993 trillion, with annual federal spending of $7.4 trillion and net interest expense consuming $1.1 trillion of tax revenue.
Ilir Xholi, Goldman Sachs’ head of swap trading, said the U.S. dollar 10-year, 10-year forward swap rate rose 47 basis points between Sept. 22 and Oct. 7, while the two-year Treasury rate was largely unchanged. Since two-year yields broadly reflect expectations for near-term Federal Reserve policy, the move pointed instead to rising term premiums—the additional compensation traders seek to hold longer-dated securities.
Xholi linked the increase to heavy September bond issuance by hyperscalers and chip companies occurring alongside substantial Treasury borrowing. A 30-year Treasury auction cleared at its highest stop-out rate since August 2000, according to the supplied data, while the two-year yield rose more than seven basis points in a single day.
Japan’s 10-year government bond yield also moved above 3%, while the report described overseas institutions reducing purchases of U.S. and European sovereign debt. Less demand from international buyers would leave governments and corporations competing more aggressively for a smaller pool of capital.
Crypto remains exposed to the risk-off trade
The funding picture also matters for cryptocurrency markets because Bitcoin has increasingly traded in line with risk-sensitive technology equities during periods of macro stress. The supplied article cited a recent 30-day correlation near 0.90 between Bitcoin and the Nasdaq, though short-term correlation measures can change rapidly.
Higher Treasury and corporate bond yields offer traders income on assets viewed as less volatile than cryptocurrencies, potentially reducing appetite for leveraged positions in Bitcoin and smaller digital tokens. The effect is most direct when falling technology shares trigger margin reductions across portfolios that include both equities and crypto assets.
The source material’s claims about stablecoin volumes and yields do not establish that stablecoins provide a shelter from market risk. Dollar-pegged tokens may reduce exposure to price swings in cryptocurrencies, but they carry issuer, reserve, platform and regulatory risks, while most stablecoins do not themselves pay interest to holders.
For the crypto market, the immediate issue is less whether AI investment continues and more whether its financing needs keep lifting the cost of capital across risk assets. A sustained rise in long-term yields, wider credit spreads and delayed data-center spending would place pressure on the technology valuations that have helped support speculative appetite across digital markets.
Rising yields reshaping AI funding? Explore how interest rates drive digital asset markets and impact leveraged tech infrastructure bets.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
