The race to build the infrastructure required for advanced artificial intelligence is escalating into a multibillion-dollar contest, with Google making a significant financial play to support one of its key competitors. According to a report by the Financial Times, citing individuals familiar with the matter, Google is preparing to provide construction loans for a colossal data center project in Texas, which is leased to AI safety firm Anthropic. The initial phase of the project, operated by Nexus Data Centers, is projected to exceed $5 billion in investment. A consortium of banks is also vying to arrange the financing, with a target to finalize deals by mid-year.
Google’s Multibillion-Dollar Bet on AI Infrastructure
The sheer scale of the planned campus underscores the capital-intensive nature of modern AI development. Situated on a sprawling 2,800-acre site, the facility is part of a broader infrastructure partnership between Google and Anthropic. Construction is already active, having secured early-stage debt financing from Eagle Point, a publicly traded closed-end investment company. The first phase aims to deliver approximately 500 megawatts of computing capacity by late 2026—enough power to supply roughly 500,000 homes. Long-term plans envision a dramatic expansion to a potential 7.7 gigawatts, a capacity that would rival that of a small city.
A critical strategic advantage of the location is its proximity to major natural gas pipelines operated by companies like Enterprise Products Partners, Energy Transfer, and Atmos Energy. This allows the project to rely on on-site gas turbines for power, a crucial factor for data centers that require immense, consistent energy supplies, especially as grid connections for such massive loads can be challenging and slow.
Strategic Context: Cloud Rivalry and AI Demands
This move highlights a complex dynamic in the tech industry. While Google Cloud competes fiercely with Microsoft Azure and Amazon Web Services to provide AI training and inference capacity, it is also directly financing infrastructure for Anthropic—a leading developer of the Claude AI models and a direct competitor to Google’s own Gemini offerings. The investment secures a major, long-term tenant for Google’s data center real estate and strengthens a pivotal partnership, ensuring Anthropic’s immense compute needs are met, likely through Google’s cloud services. It’s a strategic hedge, ensuring a key player in the AI ecosystem remains tied to Google’s infrastructure regardless of which company’s models ultimately dominate.
Federal Judge Blocks Pentagon’s Anthropic Ban
In a separate but potentially related legal development, a U.S. federal judge has temporarily halted the Pentagon’s effort to cut off federal agencies from using Anthropic’s AI tools. On Thursday, Judge Rita Lin of the U.S. District Court for the Northern District of California granted a preliminary injunction. This pauses a directive, backed by the Trump administration, that had labeled Anthropic a supply chain risk and sought to prohibit government use of its Claude chatbot.
The ruling follows a lawsuit filed by Anthropic, which argued that Defense Secretary Pete Hegseth exceeded his authority. Judge Lin described the government’s actions as “arbitrary” and expressed serious concerns about branding a U.S. company a national security threat without a clear, evidence-based legal justification. Her decision suggests the government’s actions may have been retaliatory against Anthropic for its public refusal to develop AI for lethal autonomous weapons or mass surveillance, potentially violating the company’s First Amendment rights.
The Broader Standoff Over Military AI Use
The conflict stems from a breakdown in negotiations between Anthropic and the Department of Defense. Anthropic has maintained a firm ethical stance against its models being used for offensive military applications, a position that clashed with the Pentagon’s desires. The judge’s intervention provides a significant, if temporary, reprieve for Anthropic’s federal business and sets a legal precedent challenging the executive branch’s authority to unilaterally designate domestic tech firms as security risks based on policy disagreements.
Report: Military Used Anthropic AI Despite Pentagon Ban
Adding a layer of complexity to the situation, Cointelegraph has reported that U.S. military units, including U.S. Central Command (CENTCOM), reportedly used Anthropic’s Claude AI model for operational support during a major airstrike on Iran. This alleged usage occurred even after the Trump administration’s ban order was issued, according to the report, suggesting a disconnect between the Pentagon’s public policy directives and field-level operational practices.
Source: Defense Secretary Pete Hegseth
