Key Takeaways
- • Apollo Global Management and Blackstone finalized a $35 billion private credit financing package for Anthropic, structured as a chip-backed special purpose vehicle (SPV)—one of the largest private credit transactions in history.
- • The initial deployment will add one gigawatt of AI computing capacity for Anthropic via Fluidstack starting mid-2026, with the broader partnership targeting more than 20 gigawatts across leading AI labs including OpenAI through 2028.
- • The SPV structure purchases Google AI chips, which are then leased to Anthropic, establishing a new asset-backed financing model for AI infrastructure where chips serve as collateral.
- • The deal signals a fundamental shift in AI infrastructure financing, with private capital stepping in to address acute supply constraints in computing capacity—but industry data suggests physical infrastructure delivery capacity may lag capital availability.
- • Fluidstack's selection as the infrastructure partner reflects market prioritization of deployment speed and "exceptional agility" in delivering gigawatt-scale power without extended lead times, underscoring the criticality of rapid execution in competitive AI infrastructure markets.
Summary
Apollo Global Management and Blackstone have closed a landmark $35 billion financing package designed to accelerate Anthropic's AI infrastructure expansion. The transaction, structured through a chip-backed special purpose vehicle, represents one of the largest private credit fundraisings in history and marks a pivotal moment in how artificial intelligence companies finance capital-intensive growth.
The deal mechanics center on an SPV that raises capital through a mix of debt and equity to purchase Google AI chips, which are then leased back to Anthropic. This structure creates a collateralized lending framework where semiconductor assets underpin the financing rather than traditional corporate credit metrics. The initial commitment will deliver one gigawatt of computing capacity—equivalent to powering approximately 750,000 homes—with deployment at Fluidstack-operated sites beginning in mid-2026. The broader partnership intends to enable deployment of more than 20 gigawatts of capacity across leading AI laboratories, including OpenAI, through 2028.
The selection of Fluidstack as the infrastructure partner carries strategic significance that extends beyond the transaction itself. Anthropic cited Fluidstack's "exceptional agility" in delivering gigawatt-scale power without long lead times—a criterion reflecting the market's primary constraint. As private equity and credit firms have become vital funding sources for AI companies facing acute shortages of costly and supply-constrained infrastructure, execution speed has emerged as a competitive differentiator. The ability to deploy capacity rapidly, rather than waiting for traditionally built data centers, justifies the participation of flagship capital allocators in this transaction.
However, the deal exposes a critical infrastructure bottleneck. Industry research from Turner & Townsend, reported by Capacity in November 2025, found that 83% of surveyed industry experts do not believe supply chains are equipped to deliver the advanced cooling systems required for AI-ready facilities. This finding creates tension with the multi-gigawatt commitments being announced across the industry. The fundamental market question has shifted: capital availability is no longer the constraint. Instead, the critical risk is whether physical infrastructure can absorb and deploy committed capital fast enough to meet demand.
This transaction signals a new financing era for AI infrastructure, where asset-backed structures using semiconductor equipment as collateral enable faster capital deployment than traditional corporate lending. For institutional investors, the deal demonstrates both the unprecedented capital requirements of AI development and the emerging infrastructure financing models designed to close the deployment gap.
The deal mechanics center on an SPV that raises capital through a mix of debt and equity to purchase Google AI chips, which are then leased back to Anthropic. This structure creates a collateralized lending framework where semiconductor assets underpin the financing rather than traditional corporate credit metrics. The initial commitment will deliver one gigawatt of computing capacity—equivalent to powering approximately 750,000 homes—with deployment at Fluidstack-operated sites beginning in mid-2026. The broader partnership intends to enable deployment of more than 20 gigawatts of capacity across leading AI laboratories, including OpenAI, through 2028.
The selection of Fluidstack as the infrastructure partner carries strategic significance that extends beyond the transaction itself. Anthropic cited Fluidstack's "exceptional agility" in delivering gigawatt-scale power without long lead times—a criterion reflecting the market's primary constraint. As private equity and credit firms have become vital funding sources for AI companies facing acute shortages of costly and supply-constrained infrastructure, execution speed has emerged as a competitive differentiator. The ability to deploy capacity rapidly, rather than waiting for traditionally built data centers, justifies the participation of flagship capital allocators in this transaction.
However, the deal exposes a critical infrastructure bottleneck. Industry research from Turner & Townsend, reported by Capacity in November 2025, found that 83% of surveyed industry experts do not believe supply chains are equipped to deliver the advanced cooling systems required for AI-ready facilities. This finding creates tension with the multi-gigawatt commitments being announced across the industry. The fundamental market question has shifted: capital availability is no longer the constraint. Instead, the critical risk is whether physical infrastructure can absorb and deploy committed capital fast enough to meet demand.
This transaction signals a new financing era for AI infrastructure, where asset-backed structures using semiconductor equipment as collateral enable faster capital deployment than traditional corporate lending. For institutional investors, the deal demonstrates both the unprecedented capital requirements of AI development and the emerging infrastructure financing models designed to close the deployment gap.
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.