Eagle Point Backs $1.3 Billion Loan for Anthropic's Texas AI Data Center in Major Infrastructure Deal

Eagle Point Leads $1.3 Billion Mezzanine Financing

Eagle Point Credit Management has committed approximately $1.3 billion in mezzanine financing for a massive AI data center project in Hubbard, Texas, marking one of the latest jumbo deals fueling the AI boom1

1

. The private credit loan forms part of a roughly $16 billion project-finance package for developer Nexus Data Centers, with Anthropic serving as the primary tenant2

2

. The Greenwich, Connecticut-based firm manages about $14 billion in assets and was founded in 2012 by Thomas Majewski and Stone Point Capital1

1

. Eagle Point's role as the single largest investor in the $1.3 billion loan portion stands out given its mid-sized status compared to Wall Street's largest asset managers who typically dominate jumbo private-financing markets1

1

.

Anthropic Secures Computing Power Without Building Ownership

Anthropic will anchor the 2,900-acre campus in Hubbard, located about 70 miles south of Dallas1

1

. The AI lab is racing to lock in computing power, and this deal provides access to a large new site without requiring Anthropic to build or own the infrastructure itself1

1

. The campus will feature its own gas-fired power plant to generate electricity, operating as a "behind-the-meter" power project rather than drawing from the grid1

1

. This structure aligns with Anthropic's broader strategy of leveraging outside financing to secure AI infrastructure capacity, similar to partnerships with Macquarie and others1

1

. The approach allows the Claude chatbot maker to access critical computing resources needed for training and running AI models without the upfront capital burden of data center construction1

1

.

Google Guarantees Senior Debt in Risk-Sharing Structure

Alphabet-owned Google has agreed to guarantee the senior debt portion of the deal, a move that made the financing easier to arrange2

2

. Banks including Morgan Stanley are leading the larger $15 billion senior debt package, with Eagle Point's $1.3 billion structured as mezzanine financing sitting below the safest portions2

2

. This mezzanine debt carries more risk than senior debt and typically pays higher returns to offset that risk1

1

. Google has provided similar guarantees at data centers across the United States, including a $3.2 billion lease backstop for TeraWulf's AI data center buildout2

2

. Such guarantees reassure lenders by making them more comfortable lending large sums while concentrating financial risk with fewer parties1

1

.

Rapid Deal Evolution From $150 Million to $1.3 Billion

Eagle Point began working on financing for Nexus around September 2024, initially targeting approximately $150 million in senior secured debt backed by land purchased in Hubbard2

2

. The two firms expanded and reworked the structure several times over subsequent months to reach the current deal size1

1

. Jennifer Powers, Eagle Point's head of infrastructure credit, called Nexus reaching its final investment decision in under a year "a truly remarkable achievement"1

1

. Nexus chief executive Ivan Van der Walt credited Eagle Point's team with helping complete the capital structure "at a critical moment" across four financings1

1

. The financing accelerated significantly after Anthropic won a competitive process to become the facility's primary tenant2

2

.

First Campus for Nexus Among Largest AI Infrastructure Projects

The Hubbard site represents Nexus Data Centers' first campus and will rank among the largest of dozens of data-centre projects developers are racing to bring online1

1

. Demand stems from companies training and running AI models, with Anthropic, OpenAI, Google and others signing multibillion-dollar deals to secure computing power over the past year1

1

. Nexus may issue high-yield bonds later this year to refinance the $15 billion bank-led portion, requiring the company to obtain a credit rating2

2

. This refinancing model is common in the sector, where developers typically repay construction debt by issuing new debt once projects are completed1

1

. The success of such projects depends on lenders and bond buyers remaining willing to fund AI infrastructure expansion as the AI boom continues driving unprecedented demand for computing capacity1

1

.