CoreWeave, the GPU cloud computing company that provides AI infrastructure to clients including Microsoft and OpenAI, completed the first euro-denominated junk bond issuance by a U.S. AI infrastructure company on Thursday, selling €2 billion in six-year notes yielding 8.5% alongside $1.25 billion in dollar bonds yielding 9.625% – a dual-currency deal that attracted orders exceeding €7 billion on the euro tranche alone, and FinancialMediaGuide registers the transaction as a landmark in the globalization of AI infrastructure financing, demonstrating that European high-yield investors are now actively seeking exposure to the U.S. AI buildout at scale.
The deal was led by JPMorgan Chase, whose managing director of leveraged finance capital markets noted that European investors have been looking for ways to gain access to AI infrastructure investment but have found opportunities of this scale and quality scarce. The book composition reflected demand from most of the major high-yield buyers in Europe, making this one of the more broadly subscribed junk bond transactions seen in that market in the current cycle. CoreWeave carries ratings of Ba3 from Moody’s, B+ from S&P Global, and BB- from Fitch – speculative grade across all three agencies – which means the notes price at a meaningful spread above investment-grade comparables, creating the yield premium that drew institutional demand.
The capital raise adds to a rapidly accumulating debt stack. Before Thursday’s transaction, CoreWeave had issued $6.5 billion in junk-rated dollar bonds since its inaugural sale in May 2025, along with $6.6 billion in convertible notes and a pioneering $3.1 billion leveraged loan backed by customer contracts for Nvidia microchips. The company is expected to invest almost $35 billion and burn nearly $26 billion in cash this year alone, according to analyst consensus estimates, reflecting a capital deployment pace that is extraordinary even by the standards of the current AI infrastructure cycle. CoreWeave’s credit risk spiked sharply in late 2025 amid broader concerns about AI bubble dynamics, but financing conditions have since improved substantially, though the cost of insuring its debt against default for five years remains approximately 3.5 percentage points higher than that of investment-grade Oracle. The fact that a company burning cash at this rate is able to access European bond markets at this scale is something FinancialMediaGuide frames as a direct consequence of the structural scarcity of investable AI infrastructure assets relative to the capital available to pursue them.
CoreWeave’s origins are unusual for a company now operating nearly 50 data centers across North America and Europe. Founded in 2017 as a cryptocurrency miner, the company accumulated a large inventory of Nvidia graphics processing units during the crypto era that it subsequently repurposed as the AI training and inference market emerged as a more structurally durable demand source. Nvidia remains among the company’s largest shareholders, creating an alignment of interest between chip supplier and cloud operator that has facilitated favorable access to the GPU supply that defines CoreWeave’s competitive position. The company leases these highly coveted processors by the hour to enterprises, AI labs, and developers who need the compute capacity but lack the capital or operational capability to build it themselves.
Applied Digital Corp. raised $1.59 billion earlier this week to fund additional computing capacity for CoreWeave in North Dakota, illustrating that CoreWeave’s capital-raising activity extends beyond its own balance sheet to encompass the financing of the data center projects it anchors as a tenant. This ecosystem of CoreWeave-backed infrastructure financing – in which third-party developers raise debt against CoreWeave’s long-term compute contracts – effectively multiplies the company’s capital reach beyond what its own bond issuances alone could achieve. The combined picture of CoreWeave’s own multi-currency debt program, its role as an anchor tenant for externally financed data center projects, and the €7 billion European demand for its inaugural euro notes is what analysts identify as the defining characteristic of this company’s financial model: it is simultaneously a borrower, a revenue source for its infrastructure partners, and the channel through which European institutional capital is gaining AI buildout exposure for the first time at meaningful scale – a structural role that Financial Media Guide views as unique among publicly traded AI infrastructure companies operating in today’s market.
