Pranav Hotkar
Mountain View, California - June 10, 2026 - Google is playing a central role in securing financing and infrastructure capacity for Anthropic’s rapidly expanding AI data center footprint as the AI startup moves forward with one of the industry’s largest compute expansion programs to date.
According to Reuters and multiple industry reports, Apollo Global Management and Blackstone are backing a financing structure tied to approximately USD 35 billion in new AI infrastructure capacity for Anthropic, with Google supporting key parts of the arrangement through long-term cloud, chip, and lease commitments.
The expansion is expected to add roughly one gigawatt of AI compute capacity beginning in mid-2026 at facilities managed by infrastructure provider Fluidstack. Long-term plans reportedly target more than 20 GW of compute capacity by 2028, reflecting the extraordinary scale of infrastructure now being pursued by frontier AI model developers.
Anthropic’s infrastructure strategy increasingly relies on Google’s Tensor Processing Units (TPUs) as the company seeks alternatives to NVIDIA-dominated GPU supply chains. Earlier this year, Broadcom signed a long-term agreement with Google to develop future generations of AI chips through 2031 while also enabling Anthropic to access approximately 3.5 GW of TPU-backed compute capacity beginning in 2027.
Reuters previously reported that Anthropic committed to spending approximately USD 200 billion over five years on Google Cloud services and chips, making the AI startup one of Google Cloud’s largest long-term infrastructure customers.
The latest financing initiative underscores how hyperscale AI infrastructure is increasingly being funded through complex partnerships involving cloud providers, private equity firms, chip developers, and infrastructure operators. Google has simultaneously expanded its direct investment in Anthropic, committing up to $40 billion earlier this year as competition intensifies with OpenAI, Microsoft, Amazon, and Meta.
Industry analysts say the scale of Anthropic’s infrastructure ambitions reflects a broader transition toward gigawatt-scale AI campuses designed specifically for training and inference workloads. AI systems such as Claude require massive volumes of accelerated compute, advanced networking, liquid cooling, and long-term power availability.
The agreements also highlight growing demand for custom AI silicon. Google’s TPU ecosystem, developed with Broadcom, is increasingly positioned as a lower-cost and more power-efficient alternative for large AI deployments.
As AI infrastructure investment accelerates globally, financing models are evolving alongside the technology itself. Private credit firms, hyperscale cloud providers, and infrastructure funds are now playing a major role in enabling the next generation of AI factories and high-density data center campuses.
For Google, the deepening Anthropic relationship strengthens both its cloud infrastructure business and its broader push to establish TPUs as a core platform for future AI compute deployment.
