Anthropic reportedly signs $35 billion cloud deal involving Nvidia-backed Lambda and a Texas data-center lease

Anthropic has reportedly signed a $35 billion cloud-computing agreement that is expected to be underpinned by Nvidia’s AI computing ecosystem, according to a Yahoo Finance report published Tuesday. The deal centers on long-term access to large-scale computing capacity for Anthropic’s AI workloads, with the report naming Lambda, an Nvidia-backed company, as the infrastructure partner.

The same report says Lambda holds the lease for a Texas data center that will support the arrangement. In practical terms, a data-center lease means the tenant controls access to specific facilities, and then packages that capacity into compute services for customers. For AI buyers, that can reduce the uncertainty and lead times involved in securing racks, power, cooling, and networking needed for large models.

Yahoo Finance also frames the agreement as unusually large, placing it at $35 billion. Because the company did not publish details in the report beyond the headline economics and the role of the parties, the specific term length, service levels, and expected usage quantities were not described. The report also does not disclose whether the arrangement includes exclusive capacity, multi-site redundancy, or a specific mix of Nvidia chips or servers.

The involvement of an Nvidia-backed infrastructure provider is notable because Nvidia’s GPUs and the broader “full-stack” approach around them are often used as the foundation for modern AI training and deployment. In cloud settings, customers generally buy outcomes, such as hours of compute or reserved capacity, rather than individual chips. Nvidia’s influence typically shows up through the underlying hardware in the data center and the software stack used to run neural network workloads efficiently.

Lambda’s role, as characterized in the report, highlights a recurring pattern in the AI infrastructure market. Many AI customers, particularly those scaling quickly, are looking for partners that can secure facilities and assemble the full compute environment, including power and cooling capacity, networking between servers, and the orchestration layers needed to run high-demand workloads. When those capabilities are bundled with an Nvidia-aligned supply chain, it can help buyers avoid bottlenecks that arise as demand for data-center capacity outpaces construction timelines.

Industry-wide, large cloud agreements are increasingly used to convert a company’s forecasted demand for compute into contracted capacity, which can stabilize planning and improve the odds of timely access to chips and systems. At the same time, the economics and risk allocation can differ widely depending on contract structure, including whether customers commit to minimum spend, share in cost overruns, or receive protections if performance targets are missed.

What remains unclear from the Yahoo Finance report is how the $35 billion figure is distributed across time and services. The reporting does not specify whether the total reflects a fixed commitment for reserved capacity, variable pricing tied to usage, or a mix of both. It also does not describe any geographic scope beyond the referenced Texas data center, nor does it detail whether additional capacity could be added later through expansions, new leases, or supplementary sites.

Even with the reported headline terms, investors and observers will likely focus next on confirmation from Anthropic, Lambda, or other involved parties, along with any filings or capacity announcements that could corroborate timelines. For now, the practical takeaway is that Anthropic appears to be securing very large-scale cloud compute, and the reported structure points to a data-center and infrastructure approach that routes access through Nvidia-aligned hardware supply chains and a leased Texas facility.