Inside Anthropic's $35 billion AI deal backed by Google

• Anthropic has secured approximately $35 billion in financing to support the expansion of its AI infrastructure.
• The deal revolves around Google's custom TPU chips, which will be leased to Anthropic through a specialized financing structure.
• Google and Broadcom are providing guarantees that help reduce risk for lenders and improve financing terms.
• The arrangement allows Anthropic to access large-scale computing resources without taking on the full debt directly on its balance sheet.
• The transaction highlights how major investment firms are increasingly viewing AI infrastructure as a long-term investment opportunity.

Anthropic Secures One of the Largest AI Financing Deals in the Industry

Artificial intelligence development is becoming increasingly expensive as companies race to build larger and more capable models. Training and operating advanced AI systems requires enormous amounts of computing power, specialized hardware and data center capacity, creating significant financial challenges even for some of the industry's fastest-growing firms.

Anthropic, the company behind the Claude family of AI models, has now secured approximately $35 billion in financing through a deal involving major investment firms Apollo Global Management and Blackstone. The transaction ranks among the largest private credit deals ever completed in the technology sector and reflects the growing demand for new ways to fund AI infrastructure.

Rather than providing Anthropic with a traditional corporate loan, the financing uses a structure designed specifically around the hardware needed to power the company's future AI operations.

How the Financing Structure Works

At the center of the arrangement is a special-purpose entity created to purchase Google's custom Tensor Processing Units (TPUs), which are specialized chips designed for artificial intelligence workloads.

Instead of Anthropic buying the hardware directly, the special-purpose company acquires the chips and then leases them to Anthropic for use across multiple data centers. Lease payments generated from those agreements are used to repay the investors that provided the financing.

This approach allows Anthropic to gain access to substantial computing resources without committing billions of dollars upfront to hardware purchases. It also helps the company preserve capital that can be used for research, product development and future expansion.

For lenders, the structure creates a more predictable stream of payments tied to long-term infrastructure assets rather than relying solely on the financial performance of a rapidly growing technology company.

Google and Broadcom Help Reduce Risk

A key reason the financing was able to reach such a large scale is the involvement of Google and Broadcom.

Google plays a central role by supporting lease arrangements connected to several U.S.-based data centers that will host the TPU infrastructure. The company's involvement provides additional confidence to investors given Google's established financial position and extensive cloud infrastructure business.

Broadcom, which works closely with Google on TPU development and manufacturing, is also participating through guarantees tied to portions of the financing package. These protections help reduce concerns surrounding hardware value and performance over time.

The result is a financing structure that offers investors more security than a traditional startup-focused lending arrangement. Because of those protections, portions of the debt were able to secure more favorable borrowing costs than would typically be available for a rapidly expanding AI company.

Why Anthropic Needs Massive Computing Capacity

The financing reflects a broader reality facing the AI industry: advanced models require enormous computing resources. As adoption of Claude continues to grow across enterprise and consumer markets, Anthropic must continuously expand both training and inference capacity. Training next-generation AI models can require thousands of specialized processors operating simultaneously for extended periods, while serving millions of users requires additional computing infrastructure on an ongoing basis.

The company has already expanded its partnerships with major technology providers to secure future computing capacity. Access to large-scale TPU deployments is expected to play an important role in supporting future model development and growing customer demand.

Without substantial infrastructure investments, AI developers risk falling behind competitors in a market where computing power has become one of the industry's most important strategic assets.

Wall Street Is Becoming a Major Force in AI Expansion

One of the most significant aspects of the transaction is not the size of the financing itself but who is providing it. Private credit firms such as Apollo and Blackstone have traditionally financed assets such as real estate, transportation equipment, industrial facilities and energy projects. Increasingly, they are beginning to view AI infrastructure through a similar lens.

Data centers, networking equipment and AI processors can generate long-term contractual cash flows, making them attractive to institutional investors searching for large-scale investment opportunities. Rather than viewing AI solely as a technology sector story, many investors now see the infrastructure supporting AI as an entirely new asset category.

This shift could unlock substantial pools of capital for the industry as demand for computing resources continues to accelerate.

A New Model for Funding AI Infrastructure

The Anthropic transaction may offer a glimpse into how future AI expansion will be financed. Building advanced AI systems requires more than software development. Companies must secure processors, data centers, networking infrastructure and access to reliable power sources. As costs continue to rise, alternative financing structures are becoming increasingly important.

Instead of placing all of those expenses directly on corporate balance sheets, companies can use specialized financing vehicles that spread costs across multiple parties. Hardware manufacturers, cloud providers, infrastructure operators and financial institutions can all participate in supporting large-scale AI deployments.

This approach resembles financing models commonly used in other capital-intensive industries, where long-term infrastructure assets generate predictable returns for investors while providing businesses with access to the resources they need to grow.

What It Means for the Future of AI

The deal highlights how rapidly the economics of artificial intelligence are evolving. As competition intensifies among major AI developers, access to computing power has become just as important as advances in model architecture and software capabilities. Companies that can secure large-scale infrastructure may gain a significant advantage in training future generations of AI systems.

At the same time, the growing involvement of major financial institutions suggests that AI is becoming increasingly intertwined with traditional capital markets. Investors are no longer only funding AI startups through venture capital; they are now financing the physical infrastructure that makes advanced AI possible.

The Anthropic financing package demonstrates how technology companies, chip manufacturers, cloud providers and Wall Street firms are becoming more interconnected as the AI industry expands. As demand for computing power continues to rise, similar large-scale financing arrangements are likely to become more common, helping fund the next phase of AI development across the industry.