Key Takeaways
- • Apollo Global Management and Blackstone are co-structuring a $36 billion private credit package to finance Anthropic's acquisition of Google's proprietary tensor processing units (TPUs), marking one of the largest transactions in the private credit market.
- • The financing employs a special purpose vehicle (SPV) structure in which Anthropic will lease the TPU hardware back from the acquiring entity, with long-term lease payments serving as the primary repayment source for debt holders.
- • Broadcom Inc provides residual value guarantees on the underlying hardware, effectively functioning as a credit enhancement mechanism that reduces senior lender risk and aligns the senior A1 and A2 note tranches closer to investment-grade credit profiles.
- • The transaction exemplifies an emerging trend in AI infrastructure finance where lenders underwrite hardware-backed cash flows and physical asset residual values rather than relying exclusively on corporate balance sheets and traditional credit metrics.
- • The financing is structured across multiple tranches (senior A1, A2, and subordinated B notes), with Broadcom's credit support providing direct backing to the most senior segments of the capital stack.
Summary
Apollo Global Management and Blackstone have jointly arranged a $36 billion private credit facility designed to acquire Google's custom tensor processing units (TPUs) on behalf of Anthropic, one of the largest structured financings in the private credit market. The transaction reflects sustained demand for capital to fund artificial intelligence infrastructure buildout, with both sponsors leveraging specialized structuring to address the unique risks associated with hardware-backed lending.
The financing mechanism centers on a special purpose vehicle that will purchase the TPU chips before deploying them under long-term lease arrangements with Anthropic. This approach compartmentalizes the acquisition from Anthropic's corporate credit profile, instead anchoring repayment obligations to lease cash flows generated by the AI developer. Broadcom Inc plays a critical supporting role in the transaction by providing residual value guarantees on the hardware, effectively backstopping a significant portion of the credit risk embedded in the senior debt tranches. This guarantee structure aligns the risk profile of the senior A1 and A2 note holders more closely with investment-grade credit standards, enhancing the marketability of the senior tranches to institutional investors.
The debt package is divided into multiple tranches, with senior A1 and A2 notes benefiting most directly from Broadcom's credit enhancement, while a smaller subordinated B tranche occupies a junior position in the capital structure. This tiered approach allows Apollo and Blackstone to appeal to different investor constituencies—institutional investors seeking near-investment-grade exposure through the senior notes and higher-yielding alternative investors willing to accept subordinated positions.
The transaction exemplifies a broader structural shift in AI infrastructure financing. Rather than underwriting credit based solely on corporate balance sheets and traditional leverage metrics, lenders are increasingly comfortable structuring transactions around hardware-backed cash flows, residual asset values, and third-party credit enhancements. In this case, Anthropic's lease payment obligations, the underlying value of Google's TPUs, and Broadcom's residual guarantees collectively provide multiple layers of credit support, reducing reliance on the company's standalone financial metrics. This approach has become more prevalent as demand for compute capacity outpaces traditional corporate lending appetite for AI hardware investments.
The financing mechanism centers on a special purpose vehicle that will purchase the TPU chips before deploying them under long-term lease arrangements with Anthropic. This approach compartmentalizes the acquisition from Anthropic's corporate credit profile, instead anchoring repayment obligations to lease cash flows generated by the AI developer. Broadcom Inc plays a critical supporting role in the transaction by providing residual value guarantees on the hardware, effectively backstopping a significant portion of the credit risk embedded in the senior debt tranches. This guarantee structure aligns the risk profile of the senior A1 and A2 note holders more closely with investment-grade credit standards, enhancing the marketability of the senior tranches to institutional investors.
The debt package is divided into multiple tranches, with senior A1 and A2 notes benefiting most directly from Broadcom's credit enhancement, while a smaller subordinated B tranche occupies a junior position in the capital structure. This tiered approach allows Apollo and Blackstone to appeal to different investor constituencies—institutional investors seeking near-investment-grade exposure through the senior notes and higher-yielding alternative investors willing to accept subordinated positions.
The transaction exemplifies a broader structural shift in AI infrastructure financing. Rather than underwriting credit based solely on corporate balance sheets and traditional leverage metrics, lenders are increasingly comfortable structuring transactions around hardware-backed cash flows, residual asset values, and third-party credit enhancements. In this case, Anthropic's lease payment obligations, the underlying value of Google's TPUs, and Broadcom's residual guarantees collectively provide multiple layers of credit support, reducing reliance on the company's standalone financial metrics. This approach has become more prevalent as demand for compute capacity outpaces traditional corporate lending appetite for AI hardware investments.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.