Apollo Global Management and Blackstone have finalized a massive $35 billion financing package for artificial intelligence company Anthropic, underscoring the unprecedented scale of investment flowing into AI infrastructure and data center development.
The debt package, structured across three tranches, will primarily finance Google’s custom-designed AI chips that Anthropic plans to lease as it expands its computing capabilities. The transaction ranks among the largest private credit deals ever completed and highlights how financial institutions are increasingly developing innovative financing models to support the rapidly growing artificial intelligence industry.
As AI companies continue to demand enormous computing resources, technology firms are turning to private credit markets, institutional investors, and structured financing solutions to secure the capital required for data centers, advanced semiconductors, and cloud infrastructure.
According to sources familiar with the matter, Broadcom is providing payment support for the largest senior portions of the debt package. Morgan Stanley advised Broadcom and played a key role in arranging the transaction. Additionally, investors syndicated roughly half of the $35 billion debt facility to a broader group of institutional buyers.
Representatives from Apollo, Blackstone, Anthropic, and Morgan Stanley declined to comment on the transaction, while Broadcom did not immediately respond to requests for comment.
The financing agreement represents a major milestone for the emerging chip-financing market, a sector expected to expand significantly as demand for AI infrastructure continues to accelerate worldwide. Industry observers expect this transaction to serve as a blueprint for future financing arrangements involving advanced semiconductor deployments and large-scale AI infrastructure projects.
Broadcom, which collaborates with Google on the development of Tensor Processing Units (TPUs), expects to play a central role in future AI infrastructure financing initiatives. During the company’s recent earnings call, Chief Executive Officer Hock Tan revealed that Broadcom is developing an AI infrastructure platform known as AI XPV alongside Apollo, Blackstone, and other major investors.
The initiative aims to deploy more than 20 gigawatts of compute capacity by 2028. According to Tan, Apollo is currently launching the first phase of the platform, which expects to support some of the world’s leading AI developers.
“Our strategic vision is to bring together Broadcom’s leading technology and investor partners with the strongest balance sheets to deliver at scale sufficient compute capacity at the lowest cost and power for the leading AI frontier labs, including Anthropic and OpenAI,” he said.
The financing package arrives shortly after Anthropic confidentially filed for a US initial public offering (IPO). The company appears to be positioning itself for a potential public market debut later this year as competition intensifies with rival OpenAI.
Anthropic, the developer of the Claude AI platform, recently raised $65 billion in fresh funding, bringing its valuation to approximately $965 billion, including the new investment. The company’s aggressive expansion strategy reflects the broader race among AI leaders to secure computing infrastructure capable of supporting increasingly sophisticated artificial intelligence models.
The transaction utilizes a special-purpose vehicle (SPV) structure, a financing model that has gained popularity within the AI infrastructure sector. Under this arrangement, the SPV raises debt and equity capital to purchase semiconductor assets, which are then leased to customers such as Anthropic. Debt repayment relies primarily on lease revenues generated by the chips as well as their residual value over time.
The $35 billion debt facility consists of three separate tranches. The senior portions include $6 billion in A1 notes and $24 billion in A2 notes, both of which benefit from Broadcom’s financial backing. This support enabled the debt to achieve borrowing costs aligned with Broadcom’s stronger credit profile and secure private ratings within the mid-investment-grade category.
Sources indicated that the A1 notes were sold to banks at a coupon of one percentage point above US Treasury yields. Meanwhile, the A2 tranche priced at par with a 5.75% coupon and attracted institutional investors, including Athene, Apollo’s insurance subsidiary, which typically invests in high-quality long-duration assets.
The third tranche, consisting of $4.5 billion in B notes, does not carry Broadcom’s support and therefore offers a higher coupon of 8.5%.
Beyond debt financing, Apollo’s Atlas SP Partners structured-finance division contributed approximately $800 million in equity, effectively making it the owner of the SPV established for the transaction.
One of the deal’s most innovative features involves Broadcom’s residual value support agreement. If Anthropic fails to make lease payments for a specified period, the SPV would liquidate the chips to repay investors. Should the chips fail to generate sufficient proceeds to cover investor obligations, Broadcom would compensate investors for any remaining shortfall owed to A1 and A2 noteholders.
This financing structure resembles another major AI infrastructure transaction involving Meta Platforms’ Hyperion data center project in Louisiana. In that deal, Morgan Stanley arranged a similar residual value protection mechanism, enabling investors to price the debt similarly to Meta’s corporate bonds.
The transaction highlights how the rapid expansion of artificial intelligence is reshaping global capital markets. As AI companies continue building next-generation models, financiers, technology firms, and institutional investors are collaborating to create entirely new asset classes focused on semiconductors, compute infrastructure, and data center capacity.
The $35 billion financing package for Anthropic marks a landmark moment in AI infrastructure investment and private credit markets. By bringing together Apollo, Blackstone, Broadcom, Google-linked chip infrastructure, and institutional investors, the transaction demonstrates the extraordinary capital requirements driving the artificial intelligence revolution. As demand for advanced computing power continues to surge, similar large-scale financing structures are likely to become increasingly common, fueling the next phase of AI innovation, data center expansion, and semiconductor deployment worldwide.
