- Core Scientific Inc. sold $3.3 billion in high-yield notes on April 22, 2026, in one of the largest speculative-grade debt offerings tied to AI infrastructure buildout.
- The Dallas-based company, which emerged from Chapter 11 bankruptcy in January 2024, has repositioned from Bitcoin mining to AI and high-performance computing hosting.
- High-yield bonds carry below-investment-grade credit ratings and higher coupon costs, meaning investors accept elevated default risk in exchange for greater yield.
- The raise extends a broader financing pattern in the AI infrastructure sector, where data center developers have turned to debt markets to fund capital-intensive buildouts before revenue fully matures.
What Happened
Core Scientific Inc. raised $3.3 billion through a high-yield note sale on April 22, 2026, according to Bloomberg, in what the outlet described as the latest in a series of risky borrowing transactions tied to AI infrastructure construction. The Dallas-based company has been one of the more prominent former Bitcoin miners to pivot toward AI and high-performance computing hosting, and this raise represents a significant escalation of its capital deployment.
Why It Matters
Core Scientific’s debt raise is the latest instance of AI infrastructure providers tapping speculative-grade bond markets to finance data center capacity that requires billions in upfront capital before generating revenue at scale. The company signed a 12-year, $8.7 billion contract with GPU cloud provider CoreWeave in early 2024 to convert its existing Bitcoin mining facilities into AI and HPC hosting sites — a deal that made it one of the primary infrastructure suppliers for AI model training and inference workloads. CoreWeave itself raised a $7.5 billion credit facility and completed a $1.5 billion IPO in March 2025, having similarly relied on high-leverage financing to scale before achieving profitability.
Core Scientific CEO Adam Sullivan said in the company’s February 2026 earnings call that the firm was “focused on delivering contracted megawatts on schedule” to meet its obligations under long-term AI hosting agreements, underscoring the operational pressure driving the capital raise.
Technical Details
High-yield bonds — rated below BBB- by S&P Global or Baa3 by Moody’s — compensate investors with elevated coupon rates relative to investment-grade debt, reflecting higher default probability. At $3.3 billion, this offering is among the largest single high-yield raises in the AI infrastructure sector in 2026. Core Scientific’s business model centers on power purchase agreements and long-term hosting contracts: the company controls significant electricity capacity across its data centers, which it provisions to GPU cloud providers at contracted rates. Converting legacy Bitcoin mining facilities to AI hosting requires substantial capital expenditures in cooling infrastructure, high-density power distribution, and network interconnect — costs that cannot be deferred if contracted delivery timelines are to be met.
Bitcoin mining operations typically run on power densities of around 1–3 kilowatts per rack; GPU clusters for AI training can require 30–100 kilowatts per rack, meaning physical and electrical infrastructure must be substantially rebuilt, not merely repurposed.
Who’s Affected
High-yield debt investors — including pension funds, hedge funds, and CLO managers with mandates for speculative-grade paper — are the direct counterparties, absorbing credit risk tied to Core Scientific’s construction execution and customer retention. CoreWeave, which holds multi-year, high-value contracts with Core Scientific, has an operational dependency on the company’s buildout timeline; delays in delivering contracted capacity would affect CoreWeave’s own ability to fulfill commitments to AI model developers. Other Bitcoin-mining-to-AI conversion operators — including Bit Digital, Cipher Mining, and Hut 8 — will watch bond market reception closely as a proxy for how institutional investors are pricing AI infrastructure credit risk more broadly.
What’s Next
Proceeds from the offering are expected to fund ongoing construction and expansion of Core Scientific’s AI hosting footprint across its existing data center sites. The company’s ability to service its debt obligations will depend on sustained AI compute demand and its capacity to deliver contracted megawatts on schedule without material cost overruns. Should the offering price tightly and trade well in the secondary market, it is likely to open a financing channel for other AI infrastructure operators currently exploring large-scale debt raises.