Banks plan to sell about $15 billion in debt tied to a Google-backed Anthropic data center in Texas

Banks are planning to offload about $15 billion in debt connected to a large data-center project in Texas that is backed by Google and leased to Anthropic, a report from Yahoo Finance said. The deal structure, as described in the report, would allow lenders to reduce their exposure to the financing tied to the facility.

The data center is described as being “backed by Google” and “leased to Anthropic,” placing it within the broader push by big technology companies to secure large-scale computing infrastructure for artificial intelligence workloads. In that model, data-center operators or project companies raise capital for construction and facilities, while a technology partner provides support and an AI developer signs up for long-term capacity through a lease.

While the report centers on the banks’ planned debt offload, it does not provide detailed terms in the information available here, such as the exact form of the transfer or the buyers involved. It also does not clarify whether the transaction is expected to be a full sale, a re-securitization, or another kind of balance-sheet restructuring.

Alphabet, the parent company of Google, is the key company referenced in connection with backing the project. Alphabet’s involvement, as characterized in the report, links Google’s infrastructure strategy to Anthropic’s demand for compute. For Alphabet, these kinds of arrangements can help underpin long-dated AI capacity use and reinforce its position in enterprise-scale AI infrastructure.

Anthropic, the AI company named in the report, would be the customer using the leased capacity. Lease-backed projects are common in data centers because they can give lenders and investors some visibility into revenue, which can improve the financing environment for the construction and operations phase.

The planned debt offload also points to how Wall Street is increasingly managing risk tied to AI infrastructure buildouts. Data centers are capital intensive, and large loan books tied to specific projects can become harder for banks to hold if market conditions shift or if investors demand higher returns for longer-dated exposure.

Still, major details remain undisclosed in the available account. The report does not specify the precise identity of the lenders pursuing the debt offload, the timeline for any transaction, or whether Alphabet or Anthropic would directly participate in the debt sale or take on additional guarantees.

What to watch next is whether the banks and project sponsors provide clearer information on deal mechanics, timing, and any changes to the underlying lease or financing covenants. For Alphabet and the AI sector, updates on the economics and risk allocation of these large infrastructure projects could announcement how quickly capital markets are adapting to the scale of AI demand.