Anthropic Secures Deals for Over 1 GW of Data Center Capacity as Google Weighs Financial Guarantees Ahead of IPO

SAN FRANCISCO — In a massive bid to secure the physical infrastructure necessary to power its next-generation artificial intelligence models, Anthropic has signed more than a dozen preliminary agreements to lease U.S. data centers with a combined capacity exceeding 1 gigawatt (GW).

The unprecedented scale of these agreements highlights the intensifying infrastructure war gripping the generative AI sector. According to a report by The Information, citing sources with direct knowledge of the plans, Anthropic executives have also engaged in high-level discussions with Alphabet Inc.’s Google—a primary backer—to secure financial guarantees for the lease payments.

Under the proposed arrangement, Google, which co-designs the specialized server chips Anthropic utilizes in its training and inference pipelines, would act as a financial guarantor. This safety net would reassure data center developers and landlords, mitigating the immense financial risks associated with constructing and operating facilities of this magnitude.

Chronology: Anthropic’s Rapid Rise and Capital Expansion

The data center leasing push represents the latest chapter in an extraordinary period of growth and capital accumulation for Anthropic. Founded by former OpenAI researchers, the company has rapidly positioned itself as a primary challenger—and by some metrics, the leader—in the frontier AI landscape.

[April 2026]       Alphabet announces plans to invest up to $40 Billion in Anthropic.
       │
[Late May 2026]    Anthropic raises $65 Billion in funding, reaching a $965 Billion valuation.
       │
[Early June 2026]  Anthropic confidentially files for a U.S. Initial Public Offering (IPO).
       │
[June 12, 2026]    Reports emerge of Anthropic signing >1 GW data center leases with Google guarantees.

• April 2026: Alphabet announced a monumental plan to invest up to $40 billion in Anthropic, solidifying a deep strategic partnership centered on cloud computing, hardware co-design, and model deployment.
• Late May 2026: Anthropic closed a historic funding round, raising $65 billion. The round catapulted the startup’s post-money valuation to an astonishing $965 billion, officially pushing it ahead of its chief rival, OpenAI, in private market valuation.
• Early June 2026: Capitalizing on its market momentum and surging enterprise demand, Anthropic confidentially filed for a U.S. Initial Public Offering (IPO). The company did not publicly disclose the size, pricing, or specific terms of the proposed offering.
• June 12, 2026: Details of the 1 GW data center leasing agreements and Google’s potential lease-guarantee role surfaced, illustrating the concrete infrastructure plans driving the company’s pre-IPO strategy.

Supporting Data: Understanding the Scale of 1 GW and AI Economics

To put Anthropic’s infrastructure ambitions into perspective, a 1-gigawatt (1,000-megawatt) data center footprint is roughly equivalent to the output of a modern nuclear power reactor. It is capable of powering approximately 750,000 to 1 million homes simultaneously.

For an AI company, this level of power capacity is designed to host hundreds of thousands of high-density server racks packed with advanced graphics processing units (GPUs) and specialized Tensor Processing Units (TPUs).

The Financial Mechanics of Google’s Lease Guarantees

Building data centers requires immense upfront capital expenditure (CapEx). Developers typically hesitate to construct gigawatt-scale facilities without long-term commitments from financially rock-solid tenants. Despite Anthropic’s eye-watering $965 billion valuation, it remains a pre-revenue or early-revenue startup compared to legacy tech titans.

By having Google provide a financial guarantee for the lease payments:

1. Risk Mitigation: Data center developers can secure lower-cost financing from institutional lenders, as the lease is backed by Alphabet’s AAA-grade credit profile.
2. Speed to Market: Anthropic can bypass lengthy financial vetting processes, accelerating the construction and commissioning of facilities.
3. Hardware Lock-in: The arrangement reinforces Anthropic’s reliance on Google’s custom-designed TPU chips, ensuring that the hardware running inside these facilities remains deeply integrated with Google’s ecosystem.

Official Responses and Quiet Period Constraints

When reached for comment regarding the lease agreements and the financial guarantee discussions, Google responded via email, stating that it "does not comment on rumors or speculation." Anthropic did not immediately respond to requests for comment.

Industry analysts note that Anthropic’s silence is expected. Following its confidential IPO filing earlier this month, the company is bound by strict U.S. Securities and Exchange Commission (SEC) regulations governing the "quiet period." During this window, prospective public issuers are legally restricted from making public statements or disclosing material business developments that could be perceived as hyping the company’s valuation or influencing potential investors.

Implications: The Compute War and Grid Bottlenecks

Anthropic’s massive infrastructure push carries profound implications for the broader technology sector, utility providers, and the financial markets.

1. The Power Grid Bottleneck

The sheer scale of Anthropic’s 1 GW lease agreements highlights a growing crisis in the AI industry: the availability of electric power. As AI models grow exponentially larger, the bottleneck has shifted from chip availability to power grid capacity.

Tech companies are now in a fierce race to secure grid connections, often competing directly with heavy manufacturing and local municipalities for access to electricity. Securing more than a dozen preliminary agreements across the U.S. represents a preemptive land grab by Anthropic, ensuring it has the raw power capacity locked down before competitors can claim it.

2. Shifting Hegemony in the AI Sector

For years, OpenAI was viewed as the undisputed leader of the generative AI boom. However, Anthropic’s late-May funding round at a $965 billion valuation has disrupted that narrative.

By securing massive physical infrastructure and backing it with Google’s balance sheet, Anthropic is signaling to Wall Street and enterprise clients that it possesses the operational runway and computational horsepower to sustain long-term dominance. This infrastructure is vital for maintaining the rapid development cycle of its Claude family of models, including its highly praised software development tool, Claude Code.

3. The Path to the Public Markets

The timing of these data center agreements is highly strategic. As Anthropic prepares for its IPO, prospective public investors will scrutinize its path to profitability. Running frontier AI models is an incredibly capital-intensive business, with computing costs consuming a vast portion of revenue.

By establishing these agreements and negotiating Google-backed lease guarantees, Anthropic can present a highly structured, lower-risk capital expenditure plan to IPO investors. It demonstrates that the company has solved the critical physical constraints of AI scaling—power, space, and hardware—positioning it as a mature, infrastructure-ready utility for the next industrial revolution.