Key Takeaways
- • Apollo Global Management and Blackstone are jointly financing a $35 billion AI infrastructure expansion for Anthropic, marking a significant private-equity entry into large-scale AI compute funding alongside institutional capital.
- • The initial deployment adds one gigawatt of computing capacity—equivalent to powering approximately 750,000 homes—utilizing Broadcom's custom chips and networking solutions, with infrastructure operated by Fluidstack beginning mid-2026.
- • The broader partnership targets deployment of more than 20 gigawatts of computing capacity across leading AI labs, including OpenAI, through 2028, positioning Broadcom as a strategic alternative to Nvidia for custom silicon.
- • Private-equity firms have become critical funding sources for AI infrastructure, following Meta's $27 billion financing deal with Blue Owl Capital in October, addressing acute supply constraints and capital intensity in the sector.
- • The deal reinforces industry momentum toward vertical integration and custom chip development, as major tech companies seek to reduce dependency on Nvidia while optimizing cost and power efficiency for large-scale model training.
Summary
Apollo Global Management and Blackstone have partnered to provide $35 billion in financing for Anthropic's AI computing capacity expansion using Broadcom's custom chips and networking infrastructure. The transaction, announced June 9, 2026, represents a milestone deployment in private-equity capital allocation toward large-scale AI infrastructure. The initial commitment expands Anthropic's Claude Code creator platform by one gigawatt of capacity, with physical infrastructure operated by Fluidstack at multiple data-center sites beginning in mid-2026.
The broader partnership framework targets deployment exceeding 20 gigawatts of computing capacity across multiple leading AI laboratories, including OpenAI, with full buildout scheduled through 2028. This structure positions the capital commitment as part of a platform approach rather than a single-asset transaction, enabling both asset managers to participate in scaling AI infrastructure across the ecosystem. The phased deployment model addresses both the capital intensity and execution complexity inherent in large-scale data-center infrastructure buildout.
Private-equity firms have emerged as essential funding partners for AI companies facing acute supply constraints and capital shortages. Meta's October 2025 financing of a $27 billion data-center project with Blue Owl Capital established precedent for this capital structure. The Anthropic transaction follows similar patterns, with private capital absorbing infrastructure investment risk while tech companies focus on model development and commercialization.
The partnership also validates Broadcom's competitive positioning in custom AI silicon. As major technology companies seek to reduce dependency on Nvidia's dominant GPU supply and develop proprietary chip architectures, Broadcom's custom solutions—combined with networking components—provide an integrated alternative. The deal explicitly emphasizes cost and power optimization objectives, addressing both operational economics and the escalating energy demands of frontier AI model training.
The broader partnership framework targets deployment exceeding 20 gigawatts of computing capacity across multiple leading AI laboratories, including OpenAI, with full buildout scheduled through 2028. This structure positions the capital commitment as part of a platform approach rather than a single-asset transaction, enabling both asset managers to participate in scaling AI infrastructure across the ecosystem. The phased deployment model addresses both the capital intensity and execution complexity inherent in large-scale data-center infrastructure buildout.
Private-equity firms have emerged as essential funding partners for AI companies facing acute supply constraints and capital shortages. Meta's October 2025 financing of a $27 billion data-center project with Blue Owl Capital established precedent for this capital structure. The Anthropic transaction follows similar patterns, with private capital absorbing infrastructure investment risk while tech companies focus on model development and commercialization.
The partnership also validates Broadcom's competitive positioning in custom AI silicon. As major technology companies seek to reduce dependency on Nvidia's dominant GPU supply and develop proprietary chip architectures, Broadcom's custom solutions—combined with networking components—provide an integrated alternative. The deal explicitly emphasizes cost and power optimization objectives, addressing both operational economics and the escalating energy demands of frontier AI model training.
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.