# Chip and Equipment Makers Post Strong Profits as CoreWeave Bond Yield Hits 15% By John Kim Canonical URL: https://www.tokenpost.com/news/business/34352 Published: 2026-10-18T12:00:55.000Z Updated: 2026-10-18T12:00:55.000Z Section: Business > CoreWeave generated $3.663 billion in operating cash in the first half of 2026 against $14.117 billion in cash outlays for fixed-asset purchases. Chip and equipment makers are turning artificial-intelligence demand into profit while CoreWeave faces rising borrowing costs, with a June bond's secondary-market yield reaching about 15%. CoreWeave, which rents graphics processing unit (GPU) computing capacity to large AI customers, reported second-quarter revenue of $2.575 billion, up 112% from a year earlier. Its backlog was about $104 billion at the end of June. Net interest expense totaled $640 million and the company posted a $626 million net loss. 台积电 reported third-quarter net profit of $23.9 billion, up 68.7% year over year, on revenue of $46.72 billion. Its net margin was 51.1%; 7-nanometer and more advanced processes accounted for 79% of wafer revenue, while 2-nanometer production contributed 8% as it began ramping up. ASML, a chipmaking equipment supplier, posted quarterly net profit of $5.55 billion, about 70% higher than the previous quarter, on revenue of $13.05 billion. It sold 113 new lithography systems, up from 86 in the prior quarter. Neither company's revenue comes entirely from AI, but demand for advanced chip processes and manufacturing equipment has contributed to the gains. ASML has said some customers are discussing equipment orders for 2029 and 2030, while its production plans may not meet demand in 2027 and 2028. That points to a strong pipeline for chipmaking equipment. CoreWeave still must fund GPU purchases and data-center construction before it collects revenue over the life of its multiyear customer contracts. CoreWeave's first-half cash figures show the scale of that funding need. In the first half of 2026, it generated $3.663 billion in cash from operations and made $14.117 billion in cash outlays for fixed-asset purchases, a $10.454 billion difference. Those outlays were nearly four times its operating cash flow. The roughly $104 billion backlog represents multiyear contracted business and visibility into future revenue; it is not cash already collected. GPU delivery, data-center construction and customer performance all affect how the company turns contracts into cash. The backlog can make demand more visible, but it cannot be counted as cash on hand for current expansion. CoreWeave has raised more than $35 billion this year through GPU-backed loans, high-yield bonds, convertible bonds and equity. It is seeking investment-grade financing to attract insurers and other long-term institutional investors. The company's [financing plans]() have drawn attention as it seeks funding for continued expansion. The cost of borrowing varies with the protections lenders receive. A March GPU-backed facility had a reference funding cost of about 5.9%, with its fixed rate depending on when funds were drawn. The $8.5 billion facility used specified GPUs, customer contracts and data-center leases as collateral and was structured through a separate financing entity with no general recourse to the parent company. Moody's assigned it an A3 rating, while DBRS assigned A (low). A separate CoreWeave bond issued in June had a yield of about 9.675% at issuance; its secondary-market yield later rose to about 15%. A market price decline does not directly raise the coupon on that existing fixed-rate bond. Issuing new debt in similar conditions could cost more. The bond has different credit risks and repayment terms from the GPU-backed facility, so their yields are not directly comparable. For companies building AI capacity, orders alone do not determine borrowing costs. Cash flow relative to equipment spending, contract protections, collateral and project delivery all affect how lenders price financing. ## Links in this article - [financing plans](https://www.tokenpost.com/news/business/34213)